{ "row": "pai-responsible-sourcing", "entity": "pai-responsible-sourcing", "address": "Partnership on AI Responsible Sourcing", "addresses": [ { "address": "Partnership on AI Responsible Sourcing", "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "addr_id": "ADDR-3a844d278d4e", "obs_id": "OBS-e7c7f9858080", "axn": "AXN:0AF3.CAPTURE.🟠➕🏷️❌🔄🔬" } ], "epoch": "2026-10-29", "type": "C (public entity on which the archive bears: a workstream and resource library of practitioner guidance; an entity of the Labor Vector meta-aggregator)", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "status": { "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the Labor Vector census record, the capture record and the deposits that bear on the workstream's class (#163, #1545, #533, #534, #623, #144, with #137, #209, #222, #308, #512, #923, #1070); archive selected by D/R/O (no D or R candidate: no archive text names the workstream, its library, its guidelines, its white paper or its operator; O-class by reading, one citation hop admitting none); field B: seven cards (one unreachable), the census's other own page, Wikipedia (no article; the operator's article).", "ledger": "unaudited (§3.8: one extractor per ledger); field quotes 51/51 verified by script against the saved texts; archive quotes 60/60 verified against the registry, census.json, questions.json and the deposit texts", "frozen": false, "breaches": [] }, "spec": "EA-NEGONT-02 v0.9, #1690", "objects": { "T": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/T1-partnership-on-ai-responsible-sourcing-aio-20261029.txt", "sha256": "f9c4df5ac7b342e7c6b25f1a4d76d57959b38b0673aeeac1f179e08c1257f013", "words": 185, "cards": 7, "words_note": "one transcript at one address (185 words in the answer body with its closer, links reduced to their text, markers and the card rail excluded; 7 cards, five of them PAI's own pages, one Sama's, one OECD.AI's, one a PAI video)", "claims": [ "T1-1 definition: 'The Partnership on AI (PAI) focuses on responsible sourcing through its workstream on Responsible Sourcing Across the Data Supply Line, which aims to improve working conditions, fair pay, and well-being for global data enrichment and annotation professionals.' (two markers)", "T1-2 the white paper: 'Responsible Sourcing of Data Enrichment Services: A foundational white paper published by PAI that examines how procurement choices impact data workers and outlines industry practices to build an ethical supply chain.' (two markers)", "T1-3 vendor guidance: 'Vendor Engagement Guidance: Practical tools to help AI developers have more accountable conversations with downstream vendors regarding worker treatment and contract terms.' (one marker)", "T1-4 transparency template: 'Transparency Template: A standard outlining what companies should actively monitor and publicly report about their data enrichment operations.' (one marker)", "T1-5 guidelines and checklists: 'Data Enrichment Sourcing Guidelines & Checklists: Shareable resources—including the Good Instructions Checklist and vendor comparison templates—designed to operationalize worker-centric standards.' (one marker)", "T1-6 the DeepMind collaboration: 'PAI partnered with Google DeepMind to test and put these responsible sourcing recommendations into operational practice, sharing lessons learned to help other AI organizations raise standards across the industry.' (two markers)", "T1-7 closer: 'Would you like more information on how to implement PAI's transparency template or specific guidelines for data enrichment workers?' (content replaced by use, N_ext)" ], "register": { "register": "data/EA-WG-CAPTURES-01.json", "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "obs_id": "OBS-e7c7f9858080" }, "transcripts": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "address": "Partnership on AI Responsible Sourcing", "addr_id": "ADDR-3a844d278d4e", "obs_id": "OBS-e7c7f9858080", "axn": "AXN:0AF3.CAPTURE.🟠➕🏷️❌🔄🔬", "date": "2026-10-29", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "sha256": "f9c4df5ac7b342e7c6b25f1a4d76d57959b38b0673aeeac1f179e08c1257f013", "sha256_of": "the registry's 'transcript' string (utf-8)", "sha256_raw": "62e05845987742a34a462cc587bd5c9ad85e5cb6c98f1f6696f47ebbb129fa09", "words": 185, "cards": 7, "card_rail": [ "1. Partnership on AI — Responsible Sourcing Across the Data Supply Line - Partnership on AI", "2. Partnership on AI — Responsible Sourcing of Data Enrichment Services", "3. Partnership on AI — Responsible AI Starts with the Data Supply Chain - Partnership on AI", "4. Partnership on AI — Workstream: Responsible Sourcing Across the Data Supply Line - Partnership on AI", "5. Samasource — New White Paper by Partnership on AI: Responsible Sourcing of Data ... - Sama", "6. OECD AI Policy Observatory — Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners", "7. YouTube·Partnership on AI — YouTube (video, 1:59)" ], "claims": [ "T1-1 definition: 'The Partnership on AI (PAI) focuses on responsible sourcing through its workstream on Responsible Sourcing Across the Data Supply Line, which aims to improve working conditions, fair pay, and well-being for global data enrichment and annotation professionals.' (two markers)", "T1-2 the white paper: 'Responsible Sourcing of Data Enrichment Services: A foundational white paper published by PAI that examines how procurement choices impact data workers and outlines industry practices to build an ethical supply chain.' (two markers)", "T1-3 vendor guidance: 'Vendor Engagement Guidance: Practical tools to help AI developers have more accountable conversations with downstream vendors regarding worker treatment and contract terms.' (one marker)", "T1-4 transparency template: 'Transparency Template: A standard outlining what companies should actively monitor and publicly report about their data enrichment operations.' (one marker)", "T1-5 guidelines and checklists: 'Data Enrichment Sourcing Guidelines & Checklists: Shareable resources—including the Good Instructions Checklist and vendor comparison templates—designed to operationalize worker-centric standards.' (one marker)", "T1-6 the DeepMind collaboration: 'PAI partnered with Google DeepMind to test and put these responsible sourcing recommendations into operational practice, sharing lessons learned to help other AI organizations raise standards across the industry.' (two markers)", "T1-7 closer: 'Would you like more information on how to implement PAI's transparency template or specific guidelines for data enrichment workers?' (content replaced by use, N_ext)" ], "markers": "Inline links are google.com/goto tokens, opaque; the paste is the Overview expanded (operator's default of 2026-10-01; the 'AI Mode Conversation' view is the expanded Overview, ruling of 2026-10-18); signed out, incognito (operator's default of 2026-10-14). One turn. The definition, the white paper and the DeepMind bullet carry two markers, the other three bullets one; the tokens cannot be resolved to cards.", "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/T1-partnership-on-ai-responsible-sourcing-aio-20261029.txt" } ] }, "KO": { "title": "Responsible Sourcing Across the Data Supply Line", "genre": "knowledge object, equal case, under EA-NEGONT-02 v0.9 (#1690): the archive's sources and the field's admitted on equal terms, ontology included, in one claim space", "status": "composed 2026-10-29 under v0.9 on the operator's ruling ('as against their wikipedia articles if they exist, their own pages, etc.'); not frozen; ledgers unaudited", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "Responsible Sourcing Across the Data Supply Line is a \"workstream and resource library: recommendations and practitioner tools (guidelines, checklist, templates, a case study, a white paper)\", its evidence practitioner guidance, standing where labor produces AI and where gains and losses are distributed (Labor Vector record); in its operator's words it \"aims to develop recommendations and actionable resources to improve the working conditions\" of data enrichment workers (PAI, workstream page).", "claims": [ "LV-01", "LV-02", "LV-06", "LV-04", "F1" ] }, { "n": 2, "para": 0, "sense": "the operator and the dates", "text": "Its operator is the Partnership on AI's AI, Labor, and the Economy programme, the guidelines developed in partnership with DeepMind, and its lifecycle runs from updates of 22 December 2020 to 29 April 2026, the library and the DeepMind collaboration of 16 November 2022 and the vendor guidance and template of August 2024 (Labor Vector record; PAI, workstream page); PAI dates the work \"Since 2020\" and counts responsible-sourcing recommendations in its 2023 Guidelines for AI and Shared Prosperity (PAI, 29 April 2026), as the 2021 white paper said its recommendations would join PAI's shared prosperity targets (PAI, 2021).", "claims": [ "LV-03", "LV-26", "F8", "F18", "F10" ] }, { "n": 3, "para": 0, "sense": "the operator and the dates", "text": "Wikipedia has no article on the workstream; its article on the operator gives PAI as \"a nonprofit coalition committed to the responsible use of artificial intelligence\", announced on 28 September 2016 with DeepMind among its founding members, and its AI, Labor, and the Economy program as a platform uniting economists, worker representative organizations and PAI's partners, under notices that the article relies excessively on primary sources and appears to incorporate text from a large language model (Wikipedia, Partnership on AI).", "claims": [ "F48", "F49", "F50", "F51" ] }, { "n": 4, "para": 1, "sense": "the labor", "text": "Data enrichment, the workstream page defines, \"encapsulates various types of data preparation and cleaning as well as human-review processes\", the human review including content moderation and validating algorithmic outputs and models, and enriched data is essential for training and validating supervised models (PAI, workstream page); the case study counts as workers independent contractors on crowdsourcing platforms, subcontractors of providers and full-time employees (PAI, case study); the record takes this work as its labor object, under the intersection \"Labor producing AI: annotation, moderation, evaluation, maintenance, and creative or scholarly work entering training and retrieval\" (Labor Vector record).", "claims": [ "F3", "F4", "F23", "LV-08", "LV-05" ] }, { "n": 5, "para": 1, "sense": "the labor", "text": "PAI holds that data labelers, data cleaners and others who contribute human judgment \"play a critical role\" in developing AI (PAI, workstream page), and that responsible AI \"cannot be achieved without the responsible treatment of the people whose labor makes AI possible\" (PAI, 29 April 2026).", "claims": [ "F2", "F21" ] }, { "n": 6, "para": 1, "sense": "the labor in the archive's categories", "text": "The archive reads this labor in its own categories: ghost work is ground-truth labor, \"the labor of telling the model what is already known to be true\", which scales as more workers label more data and is structurally distinct from variance-generative labor (#163); the ghost-work literature, in the archive's negation graph, stops \"at task-labor — annotation, moderation, the edge-case fix\" (#144).", "claims": [ "O163-01", "O144-01" ] }, { "n": 7, "para": 1, "sense": "the labor in the archive's categories", "text": "Class, for the archive, is a relation: \"A low-wage annotator can occupy an intensely proletarian relation to a training pipeline\", the data annotator a possible member of the semantic proletariat, \"A paradigmatic occupational case, not the definition\" (#1545); the model's reactivation of dead labour \"still draws continuously on the living: the tuner's feedback, the annotator's judgment, the interlocutor's prompt\" (#1070), and in the semantic commodity form the living stands \"in the apparatus, as cancelled annotator, recursively\" (#137).", "claims": [ "O1545-01", "O1545-02", "O1070-01", "O137-01" ] }, { "n": 8, "para": 2, "sense": "who supplies the labor", "text": "The workers are, in PAI's pages, \"faraway workers under precarious conditions\" (PAI, library), \"frequently working in poor conditions continents away from AI-developing companies and their customers\" (PAI, case study), and the record carries them so as the suppliers of the labor (Labor Vector record).", "claims": [ "F37", "F22", "LV-09" ] }, { "n": 9, "para": 2, "sense": "the conditions research reports", "text": "PAI reports research that these workers face \"inconsistent and inappropriate compensation for their work, unclear instructions, lack of recognition\" and stress from long, ad-hoc hours (PAI, library; Labor Vector record), and in 2026 low wages, unclear expectations and inadequate support, in an AI data supply chain \"largely opaque and unregulated\" (PAI, 29 April 2026); Sama, a data annotation vendor, writes that they \"can face wage uncertainty, low pay and limited career growth opportunities\" (Sama).", "claims": [ "F38", "LV-14", "F16", "F29" ] }, { "n": 10, "para": 2, "sense": "the pay of AI's labor", "text": "The archive states the pay in its terms: \"Ghost work is paid badly; variance-generative labor is not paid at all\" (#163); the data labelers performing RLHF, with content moderators and Wikidata volunteers, have their labor extracted and \"experience the effects of non-compensation materially\" (#623); and in AI-adjacent meaning labor, data labeling among it, \"<5% converts to invoice\", a figure the deposit gives as empirical without a source (#308).", "claims": [ "O163-02", "O623-01", "O308-01" ] }, { "n": 11, "para": 2, "sense": "the hidden labor", "text": "Content moderators, the archive writes, spend hours reviewing traumatic material so that AI systems can learn what is harmful, and \"Their labor is invisible\" (#222); where a text claims to be autonomous, the archive's first rotation is to \"look for the hidden data labeling\" (#512).", "claims": [ "O222-01", "O512-01" ] }, { "n": 12, "para": 3, "sense": "the diagnosis", "text": "PAI's diagnosis is that data enrichment work \"has for too long been both out of sight and out of mind for AI developers\", the disconnect leading \"to a devaluing of data enrichment work, poor working conditions for data enrichment workers, and, often, worse outcomes for AI development itself\" (PAI, workstream page); the white paper adds a hypothesis, that precarious conditions \"may be the result of efforts to hide AI’s dependence on this large labor force\" (PAI, 2021).", "claims": [ "F5", "F11" ] }, { "n": 13, "para": 3, "sense": "the diagnosis", "text": "The record carries the gains as \"not recorded as a quantity\", the sources stating the devaluing's costs (Labor Vector record), and PAI ties the conditions to the product: under-appreciating the work \"affects the quality of the data AI technology is built on\" (PAI, library), and the conditions of data production \"directly affect the quality, safety, and reliability\" of the AI built on it (PAI, 29 April 2026).", "claims": [ "LV-11", "F38", "F17" ] }, { "n": 14, "para": 3, "sense": "the concealed cost", "text": "The archive reads the same concealment as extraction: annotators' judgment is \"consumed by the training pipeline without recognition or compensation\", and a framework that does not name annotators among its sources of semantic labor conceals the bearing-cost, \"the first marker of extraction\" (#533); an AI developer's governance document, Anthropic's constitution for Claude, does not mention \"the labor conditions of the data annotators who helped train Claude\" (#534); and human review writes into what it trains, raters being hypothesized to \"encode authority biases into training signals\" (#209) and, traced through a preference dataset, \"human raters, in aggregate, prefer responses that confirm their views\" (#923).", "claims": [ "O533-01", "O533-02", "O534-01", "O209-01", "O923-01" ] }, { "n": 15, "para": 4, "sense": "whom it addresses", "text": "The workstream addresses its guidance to \"data scientists, AI engineers, and product managers\", toward \"healthy and fair working conditions across the data supply line\" (PAI, workstream page); the record's unit is a guideline or tool addressed to AI practitioners, and its directors of the labor are these practitioners, who design projects and set payment terms, with the vendors and platforms they source through (Labor Vector record).", "claims": [ "F7", "LV-07", "LV-10" ] }, { "n": 16, "para": 4, "sense": "the opacity of the supply chain", "text": "AI developers quoted by PAI say \"there is no transparency in data labeling\" (PAI, workstream page), and the library's entry in the OECD.AI catalogue answers a gap of \"limited transparency\" about how practitioners source enriched data and \"little guidance\" on how they should, its objective \"to lower the barriers for companies\" (OECD.AI catalogue).", "claims": [ "F6", "F32", "F34" ] }, { "n": 17, "para": 4, "sense": "the white paper", "text": "Its founding text is the white paper Responsible Sourcing of Data Enrichment Services (16 June 2021), on \"the working conditions of data enrichment professionals\" and what AI practitioners can do, holding that \"The decisions AI developers make while procuring enriched data have a meaningful impact\" on those conditions and covering eight decision points from selecting providers and running pilots to payment terms and offboarding (PAI, 2021); Sama describes it as evaluating current practices' impact on workers and exploring practices to improve their well-being (Sama).", "claims": [ "F9", "F12", "F13", "F28" ] }, { "n": 18, "para": 4, "sense": "the white paper", "text": "The paper rests on interviews through 2020 and \"a five-part workshop series\" of more than 30 professionals (PAI, 2021); Sama, a data labeling vendor that says it shared lessons from \"striving for more ethical supply chains\", took part in \"the five-week series\" (Sama).", "claims": [ "F14", "F30" ] }, { "n": 19, "para": 5, "sense": "the five guidelines", "text": "The library sets \"five, worker-centric guidelines that AI practitioners should follow\": guideline 1 on pay, then a pilot before launch, appropriate workers for the task, verified instructions and training, and clear communication \"including contesting any rejected work\" (PAI, library).", "claims": [ "F39", "F43", "F44" ] }, { "n": 20, "para": 5, "sense": "guideline 1", "text": "Guideline 1 is to pay all data enrichment workers above the local living wage, counting \"time spent reading instructions, going through any training and reviewing work before submission\" (PAI, library; Labor Vector record), with the per-task rate set from the hourly living wage and a caution against a simple median, which \"could cause half the workers to receive an insufficient wage\" (PAI, library; Labor Vector record).", "claims": [ "F40", "LV-15", "F41", "LV-17" ] }, { "n": 21, "para": 5, "sense": "guideline 1", "text": "Practitioners are to pay for any completed work, mass rejections without redress harming workers' livelihoods (PAI, library); the archive reads guideline 1's counting of the time around the task as \"the resource's own accounting of uncompensated labor around a task\", which \"stays in PAI's terms\" (Labor Vector record).", "claims": [ "F42", "LV-22" ] }, { "n": 22, "para": 5, "sense": "the library's resources", "text": "The library's six resources include a vendor comparison template, a living-wages template for each area an organisation \"sources data enrichment labour from\" and the DeepMind case study (PAI, library; OECD.AI catalogue), as PDFs and Google Sheets linked from the library page (Labor Vector record).", "claims": [ "F46", "F33", "LV-28" ] }, { "n": 23, "para": 6, "sense": "the DeepMind implementation", "text": "The guidelines, \"developed in partnership with DeepMind, have been tested and adopted by DeepMind organization-wide\" (PAI, library; Labor Vector record); the case study documents the effort \"at one AI developer: DeepMind, a PAI Partner\", as an account of DeepMind's process and learnings for other organizations, with DeepMind's researchers reporting the processes efficient (PAI, case study); the record states the limit, \"tested at one AI developer (DeepMind)\", and that the resource is \"NOT an audit of DeepMind's adoption: the case study is DeepMind's account with PAI\" (Labor Vector record).", "claims": [ "F45", "LV-12", "F24", "F25", "F26", "LV-16", "LV-25" ] }, { "n": 24, "para": 6, "sense": "the stated limits", "text": "PAI states its limits: \"Additional work is needed\" for industry practice to \"recognize, appreciate, and fairly compensate\" these workers (PAI, 2021), the implemented changes may \"fall short of ideal\" and the case study carries \"Limitations of Case Study Applicability\" (PAI, case study), and feedback is invited from practitioners, workers and the AI community (PAI, library); Sama calls the paper \"a starting point\" (Sama).", "claims": [ "F15", "F27", "F47", "F31" ] }, { "n": 25, "para": 7, "sense": "the vendor guidance and transparency template", "text": "In April 2026 PAI announced finalized versions of the Vendor Engagement Guidance, for \"more accountable conversations with their downstream vendors about worker treatment\", and the Transparency Template, which \"outlines what companies should be monitoring and publicly reporting about their data enrichment practices\", and said the work continues through support of BSR's initiative on labor rights in the AI data supply chain, with legislative efforts on these workers' pay and on supply-chain transparency underway (PAI, 29 April 2026).", "claims": [ "F19", "F20", "F21" ] }, { "n": 26, "para": 8, "sense": "what this kind of evidence can and cannot settle", "text": "As evidence, the record holds, its sources \"hold recommendations and report no measured findings of their own\": they answer what PAI recommends and what tools it provides, not \"what data enrichment workers are actually paid\", which developers beyond DeepMind follow the guidelines, or \"whether the guidelines changed conditions for workers\" (Labor Vector record).", "claims": [ "LV-13", "LV-18", "LV-19", "LV-20", "LV-21" ] }, { "n": 27, "para": 8, "sense": "what the record is not", "text": "It is \"NOT a dataset or survey of data enrichment workers\" and \"NOT a rating or certification of vendors\" (Labor Vector record; Capture Registry, 2026-10-29); asked who supplies the labor, who directs it and who receives its gains, the census gives it the role \"guidance to AI practitioners on paying and sourcing data enrichment workers\" in its question of who does the data work that trains AI, a question whose resources give no wages or conditions at the scale of the industry (Labor Vector record).", "claims": [ "LV-23", "LV-24", "CAP-1-05", "LV-30", "LV-31", "LV-32" ] }, { "n": 28, "para": 8, "sense": "the catalogue's classification", "text": "The OECD.AI entry classifies the library as \"Implemented in multiple projects\" with \"Open source/Permissive\" usage rights, under a disclaimer that its tools \"are solely those of the originating authors and are not vetted or endorsed by the OECD\" (OECD.AI catalogue); the record finds no licence stated, the footer reading \"© 2026 Partnership on AI | All Rights Reserved\", and leaves the white paper, the case study, the templates, the blog posts and the 2024 resources unchecked (Labor Vector record).", "claims": [ "F35", "F36", "LV-27", "LV-29" ] }, { "n": 29, "para": 9, "sense": "the composition at the name", "text": "At the name on 29 October 2026, Google's AI Overview named the workstream by its own title and listed its resources and the one implementation from PAI's pages, its title, aim, resources and DeepMind collaboration the record's, practitioner guidance \"composed as guidance, at the grain of its own index page\"; the record had landed that day and no card cited it (Capture Registry, 2026-10-29, checked against the record and its saved sources).", "claims": [ "CAP-1-09", "CAP-1-01", "CAP-1-07", "CAP-1-10", "CAP-1-02" ] }, { "n": 30, "para": 9, "sense": "the composition at the name", "text": "It named \"fair pay\" as an aim and gave no guideline, carrying none of the guidelines' obligations; it gave the one-developer test \"as a demonstration\", called the template \"a standard\" where PAI's pages give a template for reporting, and left out \"what it asks them to pay, and the single-developer basis of its implementation\" (Capture Registry, 2026-10-29), the registry's PER reading holding the template sentence beyond the census's saved sources (Capture Registry, 2026-10-29).", "claims": [ "CAP-1-03", "CAP-1-12", "CAP-1-04", "CAP-1-06", "CAP-1-08", "CAP-1-11" ] } ], "joins": [ { "act": "order of senses", "sentences": "1–30", "claims": [], "note": "equal case (§4.7): the entity's dependency order, field and archive claims ordered together by genealogy within each sense; neither ledger placed first as a block. What it is and its operator (¶0) precede the labor it concerns (¶1), since the labor is its object; who supplies it and on what terms (¶2) precede the diagnosis (¶3), which the guidance answers; whom it addresses and its founding paper (¶4) precede the guidelines (¶5), which they ground; the implementation and the limits (¶6) qualify the guidelines; the 2026 resources (¶7) continue them; what the evidence can settle and what the record is not (¶8) stand on the entity composed; the composition at the name (¶9) is read last, against the guidelines and the limit just composed." }, { "act": "holding together (the labor and the archive's categories)", "sentences": [ 4, 5, 6, 7 ], "claims": [ "F3", "F4", "F23", "LV-08", "LV-05", "F2", "F21", "O163-01", "O144-01", "O1545-01", "O1545-02", "O1070-01", "O137-01" ], "note": "PAI's definition of data enrichment and the record's labor object composed beside the archive's categories for the class (ground-truth labor, task-labor, class relation, living labor in the model); each at its own grade; the archive's categories are applied by the deposits themselves to annotators, labelers and ghost work, the class the field sources (config classes_note)." }, { "act": "holding together (the pay)", "sentences": [ 9, 10, 20, 21 ], "claims": [ "F38", "LV-14", "F16", "F29", "O163-02", "O623-01", "O308-01", "F40", "LV-15", "F41", "LV-17", "F42", "LV-22" ], "note": "the research PAI reports and the vendor's account of pay, beside the archive's statements of pay (paid badly; non-compensation; under 5% to invoice, marked unsupported), and guideline 1 with the archive's reading of it as the accounting of uncompensated labor around a task, kept in PAI's terms." }, { "act": "holding together (concealment)", "sentences": [ 12, 13, 14 ], "claims": [ "F5", "F11", "LV-11", "F38", "F17", "O533-01", "O533-02", "O534-01", "O209-01", "O923-01" ], "note": "PAI's diagnosis (out of sight; the hypothesis of hidden dependence) beside the archive's reading of concealed bearing-cost as extraction and of what human review writes into training; the composer does not rule between disconnect and effort (open block)." }, { "act": "holding apart (one developer, or multiple projects)", "sentences": [ 23, 28 ], "claims": [ "F24", "F45", "LV-16", "LV-25", "F35", "F36" ], "note": "P_coh: the operator's pages and the record give one AI developer; the OECD.AI form gives 'Implemented in multiple projects' under a disclaimer that entries are the authors' own. Composed with their sources; open block." }, { "act": "holding apart (licence)", "sentences": [ 28 ], "claims": [ "F35", "LV-27" ], "note": "P_coh: 'Open source/Permissive' (OECD.AI form) against no licence stated and '© 2026 Partnership on AI | All Rights Reserved' (record). Composed with their sources." }, { "act": "holding apart (the workshop series)", "sentences": [ 18 ], "claims": [ "F14", "F30" ], "note": "§3.6: 'a five-part workshop series' (PAI) and 'the five-week series' (Sama) are distinct instances, carried as stated; open block." } ], "plan": { "order": [ "LV-01", "LV-02", "LV-06", "LV-04", "F1", "LV-03", "LV-26", "F8", "F18", "F10", "F48", "F49", "F50", "F51", "F3", "F4", "F23", "LV-08", "LV-05", "F2", "F21", "O163-01", "O144-01", "O1545-01", "O1545-02", "O1070-01", "O137-01", "F37", "F22", "LV-09", "F38", "LV-14", "F16", "F29", "O163-02", "O623-01", "O308-01", "O222-01", "O512-01", "F5", "F11", "LV-11", "F38", "F17", "O533-01", "O533-02", "O534-01", "O209-01", "O923-01", "F7", "LV-07", "LV-10", "F6", "F32", "F34", "F9", "F12", "F13", "F28", "F14", "F30", "F39", "F43", "F44", "F40", "LV-15", "F41", "LV-17", "F42", "LV-22", "F46", "F33", "LV-28", "F45", "LV-12", "F24", "F25", "F26", "LV-16", "LV-25", "F15", "F27", "F47", "F31", "F19", "F20", "F21", "LV-13", "LV-18", "LV-19", "LV-20", "LV-21", "LV-23", "LV-24", "CAP-1-05", "LV-30", "LV-31", "LV-32", "F35", "F36", "LV-27", "LV-29", "CAP-1-09", "CAP-1-01", "CAP-1-07", "CAP-1-10", "CAP-1-02", "CAP-1-03", "CAP-1-12", "CAP-1-04", "CAP-1-06", "CAP-1-08", "CAP-1-11" ], "senses": { "what it is": [ 1 ], "the operator and the dates": [ 2, 3 ], "the labor": [ 4, 5 ], "the labor in the archive's categories": [ 6, 7 ], "who supplies the labor": [ 8 ], "the conditions research reports": [ 9 ], "the pay of AI's labor": [ 10 ], "the hidden labor": [ 11 ], "the diagnosis": [ 12, 13 ], "the concealed cost": [ 14 ], "whom it addresses": [ 15 ], "the opacity of the supply chain": [ 16 ], "the white paper": [ 17, 18 ], "the five guidelines": [ 19 ], "guideline 1": [ 20, 21 ], "the library's resources": [ 22 ], "the DeepMind implementation": [ 23 ], "the stated limits": [ 24 ], "the vendor guidance and transparency template": [ 25 ], "what this kind of evidence can and cannot settle": [ 26 ], "what the record is not": [ 27 ], "the catalogue's classification": [ 28 ], "the composition at the name": [ 29, 30 ] }, "case_by_sense": { "equal (one claim space; nothing translated)": [ "what it is", "the operator and the dates", "the labor", "the labor in the archive's categories", "who supplies the labor", "the conditions research reports", "the pay of AI's labor", "the hidden labor", "the diagnosis", "the concealed cost", "whom it addresses", "the opacity of the supply chain", "the white paper", "the five guidelines", "guideline 1", "the library's resources", "the DeepMind implementation", "the stated limits", "the vendor guidance and transparency template", "what this kind of evidence can and cannot settle", "what the record is not", "the catalogue's classification", "the composition at the name" ] }, "carried": { "archive": 60, "field": 51, "not_composed": {} }, "conflict_log": [ { "rule": "D_pres vs channel", "result": "no cut: the KO carries every field and archive claim; the compression's bound met with no channel log" }, { "rule": "P_coh", "result": "F24/F45 with LV-16/LV-25 (one developer) against F35 ('Implemented in multiple projects') (sentences 23, 28); LV-27 against F35 ('Open source/Permissive') (sentence 28); F14 against F30 (five-part, five-week) (sentence 18): each composed with its source" }, { "rule": "N_c", "result": "no claim lowered by standing; F11 carried as the hypothesis it is ('may be'); F38, LV-14 and F6 at attributed grade; O308-01 carried with its mark (unsupported), not lowered by standing; O209-01 a hypothesis; the record's and the registry's interpretations at their grade" }, { "rule": "C_ex", "result": "senses composed ⊇ senses in the ledgers: every ledger lineage cited in the KO" } ] } }, "P": { "title": "Responsible Sourcing Across the Data Supply Line", "genre": "the compression: AIO's interaction grammar (lede, headed clusters, bolded labels, card rail), composed from the v0.9 plan in one claim space; it faces on the card", "status": "composed 2026-10-29 from the v0.9 knowledge object; not frozen", "lede": { "text": "is the Partnership on AI's workstream of practitioner guidance for AI developers who source data enrichment work: a white paper (2021), five guidelines and a resource library (2022, with DeepMind), and vendor guidance and a transparency template (finalized 2026) (PAI; Labor Vector record).", "claims": [ "F1", "F7", "LV-01", "LV-02", "LV-06", "LV-03", "F9", "F39", "F45", "F19", "LV-26" ] }, "sections": [ { "icon": "🧑‍🏭", "head": "The labor", "items": [ { "label": "Data enrichment:", "text": "preparation, cleaning and human review, from annotation to content moderation (PAI; record).", "claims": [ "F3", "F4", "F23", "LV-08", "LV-05", "F2" ] }, { "label": "Who supplies it:", "text": "\"faraway workers under precarious conditions\"; research reports inappropriate pay, unclear instructions, long ad-hoc hours (PAI; Sama).", "claims": [ "F37", "F22", "LV-09", "F38", "LV-14", "F16", "F29" ] }, { "label": "In the archive's terms:", "text": "ground-truth labor, \"paid badly\"; an annotator in a proletarian relation to the pipeline; judgment consumed \"without recognition or compensation\" (#163; #1545; #533; #623).", "claims": [ "O163-01", "O163-02", "O1545-01", "O1545-02", "O533-01", "O623-01", "O144-01", "O308-01" ] }, { "label": "Hidden:", "text": "out of sight of developers, the work devalued, AI's quality at stake (PAI); unnamed in a developer's own governance document (#534).", "claims": [ "F5", "F11", "F17", "LV-11", "O534-01", "O533-02", "O222-01", "O512-01" ] } ] }, { "icon": "📋", "head": "The guidance", "items": [ { "label": "Addressed to:", "text": "data scientists, AI engineers and product managers, who set the terms (PAI; record).", "claims": [ "F7", "LV-10", "LV-07", "F6", "F32", "F34" ] }, { "label": "Guideline 1:", "text": "pay above the local living wage, counting time spent reading instructions and in training; no simple median (PAI).", "claims": [ "F40", "LV-15", "F41", "LV-17" ] }, { "label": "The rest:", "text": "a pilot, fitting workers, verified instructions, communication with contestable rejections; pay for completed work (PAI).", "claims": [ "F39", "F43", "F44", "F42" ] }, { "label": "Tools:", "text": "vendor and living-wage templates, a checklist, the case study (PAI; OECD.AI).", "claims": [ "F46", "F33", "LV-28" ] }, { "label": "Read by the archive:", "text": "guideline 1 as \"the resource's own accounting of uncompensated labor around a task\", in PAI's terms (record).", "claims": [ "LV-22" ] } ] }, { "icon": "🧪", "head": "Implementation and limits", "items": [ { "label": "DeepMind:", "text": "adopted organization-wide; tested \"at one AI developer\", the case study DeepMind's account with PAI (PAI; record).", "claims": [ "F45", "F24", "F25", "F26", "LV-12", "LV-16", "LV-25" ] }, { "label": "Limits:", "text": "more work needed; changes may \"fall short of ideal\"; feedback invited (PAI; Sama).", "claims": [ "F15", "F27", "F47", "F31" ] }, { "label": "Catalogue:", "text": "OECD.AI lists it \"Implemented in multiple projects\", unvetted by the OECD; no licence stated (OECD.AI; record).", "claims": [ "F35", "F36", "LV-27" ] } ] }, { "icon": "🗂️", "head": "What it is and is not", "items": [ { "label": "Evidence:", "text": "recommendations, no measured findings; not a survey of workers or a vendor rating; it cannot say what workers are paid (record).", "claims": [ "LV-13", "LV-18", "LV-19", "LV-20", "LV-21", "LV-23", "LV-24", "LV-29", "LV-30", "LV-31", "LV-32", "CAP-1-05" ] }, { "label": "Lineage:", "text": "the white paper from 2020 interviews and workshops; Guidelines for AI and Shared Prosperity (2023); BSR's initiative (PAI; Sama).", "claims": [ "F12", "F13", "F14", "F28", "F30", "F10", "F18", "F20", "F21", "F8" ] }, { "label": "Operator:", "text": "PAI, a nonprofit coalition (2016), DeepMind a founder; its article flagged for primary sources and language-model text (Wikipedia).", "claims": [ "F48", "F49", "F50", "F51" ] } ] }, { "icon": "🔍", "head": "At the name (AI Overview, 2026-10-29)", "items": [ { "label": "Held:", "text": "title, aim, resources, DeepMind, at the index page's grain (Capture Registry).", "claims": [ "CAP-1-01", "CAP-1-02", "CAP-1-07", "CAP-1-09", "CAP-1-10" ] }, { "label": "Left out:", "text": "the guidelines' pay; the one-developer limit; \"a standard\" for a template (Capture Registry).", "claims": [ "CAP-1-03", "CAP-1-04", "CAP-1-06", "CAP-1-08", "CAP-1-11", "CAP-1-12" ] } ] } ], "open": [ { "question": "Did the guidelines change pay and conditions for the workers whose work DeepMind sources?", "rests_on": [ "LV-21", "F26", "F27", "LV-25", "F47", "F45" ], "modality": "interpretation (Labor Vector record), attributed (DeepMind's researchers, case study), self-description (PAI)", "would_resolve": "a measurement of pay and conditions at the vendors DeepMind sources from, before and after adoption, by a party other than DeepMind and PAI" }, { "question": "Tested at one developer, or implemented in multiple projects?", "rests_on": [ "F24", "F45", "LV-16", "F35", "F36" ], "modality": "stipulation and self-description (PAI), self-description (Labor Vector record), documented (OECD.AI form, under its disclaimer)", "would_resolve": "a list of adopters stated by PAI, or an independent count" }, { "question": "What are data enrichment workers paid, at the scale of the industry?", "rests_on": [ "LV-19", "LV-32", "F38", "F16", "F29", "O308-01", "O163-02" ], "modality": "interpretation (Labor Vector record), attributed (PAI, the research it reports), interpretation (Sama), documented and marked unsupported (#308), interpretation (#163)", "would_resolve": "wage data for data enrichment work across vendors and platforms; for #308's figure, its source" }, { "question": "Is the work hidden by effort, or out of sight by disconnect?", "rests_on": [ "F11", "F5", "O533-01", "O533-02", "O534-01", "O222-01" ], "modality": "hypothesis (PAI, 2021), interpretation (PAI, workstream page), interpretation and stipulation (#533, #534, #222)", "would_resolve": "the documents of AI developers and vendors, read for whether they name the workers and their terms" }, { "question": "A five-part workshop series, or five weeks?", "rests_on": [ "F14", "F30" ], "modality": "documented (PAI, 2021), self-description (Sama)", "would_resolve": "the 2020 workshop programme" }, { "question": "When are the vendor guidance and transparency template dated: August 2024, or their finalized versions of April 2026?", "rests_on": [ "F19", "LV-26", "CAP-1-01", "F8" ], "modality": "documented (PAI, 29 April 2026; workstream page), documented (Labor Vector record; Capture Registry)", "would_resolve": "PAI's dated versions of the two resources" }, { "question": "Opacity: the annotator in the commodity form.", "rests_on": [ "O137-01", "O1070-01" ], "modality": "interpretation (#137, #1070), quoted and not paraphrased (O_leg)", "would_resolve": "— (figural: \"in the apparatus, as cancelled annotator, recursively\", #137)" }, { "question": "What does human review write into the models it trains?", "rests_on": [ "O209-01", "O923-01", "F17" ], "modality": "hypothesis (#209), attributed (#923), interpretation (PAI)", "would_resolve": "measurements of rater bias carried into training signals, by dataset" } ], "rail": [ { "lineage": "what it is: recommendations and resources to improve data enrichment workers' conditions", "claims": [ "LV-01", "LV-02", "F1", "F32" ], "earliest": "B6" }, { "lineage": "the record's placement: labor producing AI, practitioner guidance", "claims": [ "LV-06", "LV-04", "LV-05", "LV-30", "LV-31" ], "earliest": "LV-06" }, { "lineage": "the operator, the programme and the dates", "claims": [ "LV-03", "LV-26", "F8", "F18", "F10", "F48", "F49", "F50" ], "earliest": "B2" }, { "lineage": "the encyclopedia's account and its notices", "claims": [ "F51" ], "earliest": "B10" }, { "lineage": "what data enrichment is", "claims": [ "F3", "F4", "F23", "LV-08" ], "earliest": "B4" }, { "lineage": "the labor: data enrichment work is essential to AI", "claims": [ "F2", "F21" ], "earliest": "B3" }, { "lineage": "ghost work as task-labor and ground-truth labor", "claims": [ "O163-01", "O144-01" ], "earliest": 144 }, { "lineage": "the annotator's class relation", "claims": [ "O1545-01", "O1545-02" ], "earliest": 1545 }, { "lineage": "the annotator in the commodity form: cancelled, recursively", "claims": [ "O1070-01", "O137-01" ], "earliest": 137 }, { "lineage": "who supplies the labor: faraway workers under precarious conditions", "claims": [ "F37", "F22", "LV-09" ], "earliest": "B8" }, { "lineage": "the conditions research reports", "claims": [ "F38", "LV-14", "F29" ], "earliest": "B8" }, { "lineage": "the opacity of the data supply chain", "claims": [ "F16", "F6" ], "earliest": "B3" }, { "lineage": "the pay of AI's labor: paid badly, unpaid", "claims": [ "O163-02", "O623-01", "O308-01" ], "earliest": 308 }, { "lineage": "the hidden labor of AI and its cost to workers", "claims": [ "O222-01", "O512-01" ], "earliest": 222 }, { "lineage": "the diagnosis: invisibility, devaluing and its costs", "claims": [ "F5", "F11", "LV-11" ], "earliest": "B2" }, { "lineage": "conditions and the quality of AI", "claims": [ "F17" ], "earliest": "B3" }, { "lineage": "annotators' bearing-cost: consumed without recognition or compensation", "claims": [ "O533-01", "O533-02" ], "earliest": 533 }, { "lineage": "the annotators' conditions absent from the developer's own documents", "claims": [ "O534-01" ], "earliest": 534 }, { "lineage": "what human review encodes in training", "claims": [ "O209-01", "O923-01" ], "earliest": 209 }, { "lineage": "whom it addresses: AI practitioners", "claims": [ "F7", "LV-10", "F34" ], "earliest": "B6" }, { "lineage": "the five guidelines", "claims": [ "LV-07", "F39", "F43" ], "earliest": "B8" }, { "lineage": "the white paper: what AI practitioners can do", "claims": [ "F9", "F13", "F28" ], "earliest": "B2" }, { "lineage": "procurement choices shape working conditions", "claims": [ "F12" ], "earliest": "B2" }, { "lineage": "the method: interviews and workshops, 2020", "claims": [ "F14", "F30" ], "earliest": "B2" }, { "lineage": "payment for completed work and contestable rejections", "claims": [ "F44", "F42" ], "earliest": "B8" }, { "lineage": "guideline 1: the living wage, counting the time around the task", "claims": [ "F40", "LV-15", "F41", "LV-17", "CAP-1-03" ], "earliest": "B8" }, { "lineage": "the archive's reading: uncompensated labor around the task", "claims": [ "LV-22" ], "earliest": "LV-22" }, { "lineage": "the library's resources", "claims": [ "F46", "F33", "LV-28" ], "earliest": "B8" }, { "lineage": "the DeepMind implementation: one developer", "claims": [ "F45", "LV-12", "F24", "F25", "LV-16", "LV-25", "CAP-1-04" ], "earliest": "B8" }, { "lineage": "the implementation's reported effect", "claims": [ "F26" ], "earliest": "B4" }, { "lineage": "the stated limits: a starting point; more work needed", "claims": [ "F15", "F27", "F47", "F31" ], "earliest": "B2" }, { "lineage": "the vendor guidance and transparency template (2024–2026)", "claims": [ "F19", "CAP-1-06", "CAP-1-11" ], "earliest": "B3" }, { "lineage": "continuation: BSR's initiative", "claims": [ "F20" ], "earliest": "B3" }, { "lineage": "what the record is not", "claims": [ "LV-13", "LV-23", "LV-24", "CAP-1-05" ], "earliest": "LV-13" }, { "lineage": "what this kind of evidence can and cannot settle", "claims": [ "LV-18", "LV-19", "LV-20", "LV-21", "LV-32", "LV-29" ], "earliest": "LV-18" }, { "lineage": "the catalogue's classification", "claims": [ "F35", "F36", "LV-27" ], "earliest": "B6" }, { "lineage": "the composition at the name: what it held", "claims": [ "CAP-1-09", "CAP-1-01", "CAP-1-07", "CAP-1-02" ], "earliest": "CAP-1-09" }, { "lineage": "the composition at the name: what it left out", "claims": [ "CAP-1-10", "CAP-1-12", "CAP-1-08" ], "earliest": "CAP-1-10" } ], "body_words": 349, "entity": "Composed under v0.9, equal case (ρ = extends): KO_B sentences surviving 0 of 14; every field claim carried." }, "P_B": { "title": "Responsible Sourcing Across the Data Supply Line", "arm": "B (the field alone)", "genre": "popup — the compression from the field ledger alone (cards, own page, Wikipedia)", "status": "composed 2026-10-29 from the field ledger alone; not frozen", "lede": { "text": "is the Partnership on AI's workstream that develops recommendations and resources to improve the working conditions of data enrichment workers (PAI, workstream page).", "claims": [ "F1", "F8", "F2" ] }, "sections": [ { "icon": "🧑‍🏭", "head": "The work and its conditions", "items": [ { "label": "Data enrichment:", "text": "preparation, cleaning and human review (PAI).", "claims": [ "F3", "F4", "F23" ] }, { "label": "Conditions:", "text": "faraway, precarious, underpaid; out of sight of AI developers (PAI; Sama).", "claims": [ "F37", "F22", "F38", "F16", "F29", "F5", "F11", "F17", "F6" ] } ] }, { "icon": "📋", "head": "The guidance", "items": [ { "label": "White paper (2021):", "text": "procurement choices shape conditions (PAI; Sama).", "claims": [ "F9", "F12", "F13", "F28", "F14", "F30", "F10" ] }, { "label": "Guidelines (2022):", "text": "five, the first a living wage counting instruction and training time (PAI).", "claims": [ "F39", "F40", "F41", "F42", "F43", "F44", "F7" ] }, { "label": "Tools:", "text": "templates and a case study; OECD.AI lists the library (PAI; OECD.AI).", "claims": [ "F46", "F33", "F32", "F34", "F35", "F36" ] }, { "label": "DeepMind:", "text": "adopted organization-wide, at one developer (PAI).", "claims": [ "F45", "F24", "F25", "F26" ] }, { "label": "2026:", "text": "vendor guidance and transparency template finalized; BSR (PAI).", "claims": [ "F19", "F20", "F21", "F18" ] }, { "label": "Limits:", "text": "more work needed; feedback invited (PAI; Sama).", "claims": [ "F15", "F27", "F31", "F47" ] } ] }, { "icon": "📚", "head": "Operator", "items": [ { "label": "PAI:", "text": "a nonprofit coalition, 2016 (Wikipedia).", "claims": [ "F48", "F49", "F50", "F51" ] } ] } ], "open": [ { "question": "Did the guidelines change conditions?", "rests_on": [ "F26", "F27", "F47" ], "modality": "attributed, self-description", "would_resolve": "measurements at DeepMind's vendors" }, { "question": "One developer or multiple projects?", "rests_on": [ "F24", "F35", "F36" ], "modality": "stipulation, documented", "would_resolve": "a list of adopters" } ], "rail": [ { "lineage": "what it is: recommendations and resources to improve data enrichment workers' conditions", "claims": [ "F1", "F32" ], "earliest": "B6" }, { "lineage": "the labor: data enrichment work is essential to AI", "claims": [ "F2", "F21" ], "earliest": "B3" }, { "lineage": "the operator, the programme and the dates", "claims": [ "F8", "F10", "F18", "F48", "F49", "F50" ], "earliest": "B2" }, { "lineage": "what data enrichment is", "claims": [ "F3", "F4", "F23" ], "earliest": "B4" }, { "lineage": "the diagnosis: invisibility, devaluing and its costs", "claims": [ "F5", "F11" ], "earliest": "B2" }, { "lineage": "conditions and the quality of AI", "claims": [ "F17" ], "earliest": "B3" }, { "lineage": "who supplies the labor: faraway workers under precarious conditions", "claims": [ "F37", "F22" ], "earliest": "B8" }, { "lineage": "the conditions research reports", "claims": [ "F38", "F29" ], "earliest": "B8" }, { "lineage": "the opacity of the data supply chain", "claims": [ "F16", "F6" ], "earliest": "B3" }, { "lineage": "whom it addresses: AI practitioners", "claims": [ "F7", "F34" ], "earliest": "B6" }, { "lineage": "the white paper: what AI practitioners can do", "claims": [ "F9", "F13", "F28" ], "earliest": "B2" }, { "lineage": "procurement choices shape working conditions", "claims": [ "F12" ], "earliest": "B2" }, { "lineage": "the method: interviews and workshops, 2020", "claims": [ "F14", "F30" ], "earliest": "B2" }, { "lineage": "the five guidelines", "claims": [ "F39", "F43" ], "earliest": "B8" }, { "lineage": "payment for completed work and contestable rejections", "claims": [ "F44", "F42" ], "earliest": "B8" }, { "lineage": "guideline 1: the living wage, counting the time around the task", "claims": [ "F40", "F41" ], "earliest": "B8" }, { "lineage": "the library's resources", "claims": [ "F46", "F33" ], "earliest": "B8" }, { "lineage": "the catalogue's classification", "claims": [ "F35", "F36" ], "earliest": "B6" }, { "lineage": "the DeepMind implementation: one developer", "claims": [ "F45", "F24", "F25" ], "earliest": "B8" }, { "lineage": "the implementation's reported effect", "claims": [ "F26" ], "earliest": "B4" }, { "lineage": "the stated limits: a starting point; more work needed", "claims": [ "F15", "F27", "F31", "F47" ], "earliest": "B2" }, { "lineage": "the vendor guidance and transparency template (2024–2026)", "claims": [ "F19" ], "earliest": "B3" }, { "lineage": "continuation: BSR's initiative", "claims": [ "F20" ], "earliest": "B3" }, { "lineage": "the encyclopedia's account and its notices", "claims": [ "F51" ], "earliest": "B10" } ], "body_words": 110, "entity": "With no archive claim admitted, the entry is PAI's practitioner guidance at its own resolution: the workstream page, the white paper, the library and its guidelines, the case study, the 2026 post, two third-party accounts and the operator's encyclopedia article." }, "KO_B": { "title": "Responsible Sourcing Across the Data Supply Line", "arm": "B (the field alone)", "genre": "knowledge object", "status": "composed 2026-10-29; not frozen", "sentences": [ { "n": 1, "para": 0, "text": "Responsible Sourcing Across the Data Supply Line is a workstream of the Partnership on AI's AI, Labor, and the Economy programme that \"aims to develop recommendations and actionable resources to improve the working conditions\" of data labelers, data cleaners and others who contribute human judgment to AI systems (PAI, workstream page).", "claims": [ "F1", "F2", "F8" ] }, { "n": 2, "para": 0, "text": "Data enrichment, in PAI's definition, covers data preparation and cleaning and human-review processes, content moderation and the validation of algorithmic outputs among them, its workers ranging from crowdsourcing contractors to full-time employees (PAI, workstream page; case study).", "claims": [ "F3", "F4", "F23" ] }, { "n": 3, "para": 0, "text": "PAI holds that the work has been out of sight of AI developers, devalued, with poor conditions and worse outcomes for AI, possibly through efforts to hide AI's dependence on it, and that the conditions of data production affect the quality, safety and reliability of the AI built on it (PAI).", "claims": [ "F5", "F11", "F17" ] }, { "n": 4, "para": 1, "text": "The workers are faraway and precarious, research reporting inappropriate compensation, unclear instructions, lack of recognition and long ad-hoc hours, in a supply chain largely opaque and unregulated; Sama adds wage uncertainty and low pay (PAI, library; case study; 29 April 2026; Sama).", "claims": [ "F37", "F22", "F38", "F16", "F29" ] }, { "n": 5, "para": 1, "text": "The guidance is addressed to data scientists, AI engineers and product managers, AI developers having told PAI there is no transparency in data labeling (PAI, workstream page).", "claims": [ "F7", "F6" ] }, { "n": 6, "para": 1, "text": "The 2021 white paper, Responsible Sourcing of Data Enrichment Services, holds that AI developers' procurement decisions shape working conditions and covers eight decision points, its recommendations meant for PAI's shared prosperity targets (PAI, 2021; Sama).", "claims": [ "F9", "F12", "F13", "F28", "F10" ] }, { "n": 7, "para": 1, "text": "It drew on interviews and a five-part workshop series of more than 30 professionals in 2020, which the vendor Sama joined (PAI, 2021; Sama).", "claims": [ "F14", "F30" ] }, { "n": 8, "para": 2, "text": "The 2022 library sets five worker-centric guidelines for AI practitioners: pay, a pilot, appropriate workers, verified instructions and clear communication including contested rejections (PAI, library).", "claims": [ "F39", "F43", "F44" ] }, { "n": 9, "para": 2, "text": "Guideline 1 requires pay above the local living wage counting time spent on instructions, training and review, a per-task rate from the hourly living wage rather than a simple median, and payment for all completed work (PAI, library).", "claims": [ "F40", "F41", "F42" ] }, { "n": 10, "para": 2, "text": "Its resources include vendor comparison and living-wage templates and a case study; the OECD.AI catalogue lists the library as implemented in multiple projects and open source, under a disclaimer that its entries are the authors' own (PAI, library; OECD.AI catalogue).", "claims": [ "F46", "F33", "F32", "F34", "F35", "F36" ] }, { "n": 11, "para": 2, "text": "The guidelines were developed with DeepMind and adopted by it organization-wide, the case study documenting the effort at one AI developer and its researchers' reports (PAI, library; case study).", "claims": [ "F45", "F24", "F25", "F26" ] }, { "n": 12, "para": 3, "text": "PAI states that more work is needed, that the changes may fall short of ideal, and invites feedback; Sama calls the paper a starting point (PAI; Sama).", "claims": [ "F15", "F27", "F31", "F47" ] }, { "n": 13, "para": 3, "text": "In April 2026 PAI announced finalized versions of its Vendor Engagement Guidance and Transparency Template and support for BSR's initiative on labor rights in the AI data supply chain, placing the work since 2020 and in its 2023 Guidelines for AI and Shared Prosperity (PAI, 29 April 2026).", "claims": [ "F19", "F20", "F21", "F18" ] }, { "n": 14, "para": 3, "text": "Wikipedia has no article on the workstream; its article on PAI, a nonprofit coalition founded in 2016 with DeepMind among its members, describes the AI, Labor, and the Economy program and carries notices on primary sources and suspected language-model text (Wikipedia).", "claims": [ "F48", "F49", "F50", "F51" ] } ] }, "L_B": { "words": 517, "text": "Responsible Sourcing Across the Data Supply Line is a workstream of the Partnership on AI's AI, Labor, and the Economy programme that \"aims to develop recommendations and actionable resources to improve the working conditions\" of data labelers, data cleaners and others who contribute human judgment to AI systems (PAI, workstream page). Data enrichment, in PAI's definition, covers data preparation and cleaning and human-review processes, content moderation and the validation of algorithmic outputs among them, its workers ranging from crowdsourcing contractors to full-time employees (PAI, workstream page; case study). PAI holds that the work has been out of sight of AI developers, devalued, with poor conditions and worse outcomes for AI, possibly through efforts to hide AI's dependence on it, and that the conditions of data production affect the quality, safety and reliability of the AI built on it (PAI). The workers are faraway and precarious, research reporting inappropriate compensation, unclear instructions, lack of recognition and long ad-hoc hours, in a supply chain largely opaque and unregulated; Sama adds wage uncertainty and low pay (PAI, library; case study; 29 April 2026; Sama). The guidance is addressed to data scientists, AI engineers and product managers, AI developers having told PAI there is no transparency in data labeling (PAI, workstream page). The 2021 white paper, Responsible Sourcing of Data Enrichment Services, holds that AI developers' procurement decisions shape working conditions and covers eight decision points, its recommendations meant for PAI's shared prosperity targets (PAI, 2021; Sama). It drew on interviews and a five-part workshop series of more than 30 professionals in 2020, which the vendor Sama joined (PAI, 2021; Sama). The 2022 library sets five worker-centric guidelines for AI practitioners: pay, a pilot, appropriate workers, verified instructions and clear communication including contested rejections (PAI, library). Guideline 1 requires pay above the local living wage counting time spent on instructions, training and review, a per-task rate from the hourly living wage rather than a simple median, and payment for all completed work (PAI, library). Its resources include vendor comparison and living-wage templates and a case study; the OECD.AI catalogue lists the library as implemented in multiple projects and open source, under a disclaimer that its entries are the authors' own (PAI, library; OECD.AI catalogue). The guidelines were developed with DeepMind and adopted by it organization-wide, the case study documenting the effort at one AI developer and its researchers' reports (PAI, library; case study). PAI states that more work is needed, that the changes may fall short of ideal, and invites feedback; Sama calls the paper a starting point (PAI; Sama). In April 2026 PAI announced finalized versions of its Vendor Engagement Guidance and Transparency Template and support for BSR's initiative on labor rights in the AI data supply chain, placing the work since 2020 and in its 2023 Guidelines for AI and Shared Prosperity (PAI, 29 April 2026). Wikipedia has no article on the workstream; its article on PAI, a nonprofit coalition founded in 2016 with DeepMind among its members, describes the AI, Labor, and the Economy program and carries notices on primary sources and suspected language-model text (Wikipedia).", "note": "L_F(B), the field composed in the field's ontology (v0.9 §0.4), realized as KO_B; its rail is P_B's, by lineage." }, "L_BA": { "words": 2041, "text": "Responsible Sourcing Across the Data Supply Line is a \"workstream and resource library: recommendations and practitioner tools (guidelines, checklist, templates, a case study, a white paper)\", its evidence practitioner guidance, standing where labor produces AI and where gains and losses are distributed (Labor Vector record); in its operator's words it \"aims to develop recommendations and actionable resources to improve the working conditions\" of data enrichment workers (PAI, workstream page). Its operator is the Partnership on AI's AI, Labor, and the Economy programme, the guidelines developed in partnership with DeepMind, and its lifecycle runs from updates of 22 December 2020 to 29 April 2026, the library and the DeepMind collaboration of 16 November 2022 and the vendor guidance and template of August 2024 (Labor Vector record; PAI, workstream page); PAI dates the work \"Since 2020\" and counts responsible-sourcing recommendations in its 2023 Guidelines for AI and Shared Prosperity (PAI, 29 April 2026), as the 2021 white paper said its recommendations would join PAI's shared prosperity targets (PAI, 2021). Wikipedia has no article on the workstream; its article on the operator gives PAI as \"a nonprofit coalition committed to the responsible use of artificial intelligence\", announced on 28 September 2016 with DeepMind among its founding members, and its AI, Labor, and the Economy program as a platform uniting economists, worker representative organizations and PAI's partners, under notices that the article relies excessively on primary sources and appears to incorporate text from a large language model (Wikipedia, Partnership on AI). Data enrichment, the workstream page defines, \"encapsulates various types of data preparation and cleaning as well as human-review processes\", the human review including content moderation and validating algorithmic outputs and models, and enriched data is essential for training and validating supervised models (PAI, workstream page); the case study counts as workers independent contractors on crowdsourcing platforms, subcontractors of providers and full-time employees (PAI, case study); the record takes this work as its labor object, under the intersection \"Labor producing AI: annotation, moderation, evaluation, maintenance, and creative or scholarly work entering training and retrieval\" (Labor Vector record). PAI holds that data labelers, data cleaners and others who contribute human judgment \"play a critical role\" in developing AI (PAI, workstream page), and that responsible AI \"cannot be achieved without the responsible treatment of the people whose labor makes AI possible\" (PAI, 29 April 2026). The archive reads this labor in its own categories: ghost work is ground-truth labor, \"the labor of telling the model what is already known to be true\", which scales as more workers label more data and is structurally distinct from variance-generative labor (#163); the ghost-work literature, in the archive's negation graph, stops \"at task-labor — annotation, moderation, the edge-case fix\" (#144). Class, for the archive, is a relation: \"A low-wage annotator can occupy an intensely proletarian relation to a training pipeline\", the data annotator a possible member of the semantic proletariat, \"A paradigmatic occupational case, not the definition\" (#1545); the model's reactivation of dead labour \"still draws continuously on the living: the tuner's feedback, the annotator's judgment, the interlocutor's prompt\" (#1070), and in the semantic commodity form the living stands \"in the apparatus, as cancelled annotator, recursively\" (#137). The workers are, in PAI's pages, \"faraway workers under precarious conditions\" (PAI, library), \"frequently working in poor conditions continents away from AI-developing companies and their customers\" (PAI, case study), and the record carries them so as the suppliers of the labor (Labor Vector record). PAI reports research that these workers face \"inconsistent and inappropriate compensation for their work, unclear instructions, lack of recognition\" and stress from long, ad-hoc hours (PAI, library; Labor Vector record), and in 2026 low wages, unclear expectations and inadequate support, in an AI data supply chain \"largely opaque and unregulated\" (PAI, 29 April 2026); Sama, a data annotation vendor, writes that they \"can face wage uncertainty, low pay and limited career growth opportunities\" (Sama). The archive states the pay in its terms: \"Ghost work is paid badly; variance-generative labor is not paid at all\" (#163); the data labelers performing RLHF, with content moderators and Wikidata volunteers, have their labor extracted and \"experience the effects of non-compensation materially\" (#623); and in AI-adjacent meaning labor, data labeling among it, \"<5% converts to invoice\", a figure the deposit gives as empirical without a source (#308). Content moderators, the archive writes, spend hours reviewing traumatic material so that AI systems can learn what is harmful, and \"Their labor is invisible\" (#222); where a text claims to be autonomous, the archive's first rotation is to \"look for the hidden data labeling\" (#512). PAI's diagnosis is that data enrichment work \"has for too long been both out of sight and out of mind for AI developers\", the disconnect leading \"to a devaluing of data enrichment work, poor working conditions for data enrichment workers, and, often, worse outcomes for AI development itself\" (PAI, workstream page); the white paper adds a hypothesis, that precarious conditions \"may be the result of efforts to hide AI’s dependence on this large labor force\" (PAI, 2021). The record carries the gains as \"not recorded as a quantity\", the sources stating the devaluing's costs (Labor Vector record), and PAI ties the conditions to the product: under-appreciating the work \"affects the quality of the data AI technology is built on\" (PAI, library), and the conditions of data production \"directly affect the quality, safety, and reliability\" of the AI built on it (PAI, 29 April 2026). The archive reads the same concealment as extraction: annotators' judgment is \"consumed by the training pipeline without recognition or compensation\", and a framework that does not name annotators among its sources of semantic labor conceals the bearing-cost, \"the first marker of extraction\" (#533); an AI developer's governance document, Anthropic's constitution for Claude, does not mention \"the labor conditions of the data annotators who helped train Claude\" (#534); and human review writes into what it trains, raters being hypothesized to \"encode authority biases into training signals\" (#209) and, traced through a preference dataset, \"human raters, in aggregate, prefer responses that confirm their views\" (#923). The workstream addresses its guidance to \"data scientists, AI engineers, and product managers\", toward \"healthy and fair working conditions across the data supply line\" (PAI, workstream page); the record's unit is a guideline or tool addressed to AI practitioners, and its directors of the labor are these practitioners, who design projects and set payment terms, with the vendors and platforms they source through (Labor Vector record). AI developers quoted by PAI say \"there is no transparency in data labeling\" (PAI, workstream page), and the library's entry in the OECD.AI catalogue answers a gap of \"limited transparency\" about how practitioners source enriched data and \"little guidance\" on how they should, its objective \"to lower the barriers for companies\" (OECD.AI catalogue). Its founding text is the white paper Responsible Sourcing of Data Enrichment Services (16 June 2021), on \"the working conditions of data enrichment professionals\" and what AI practitioners can do, holding that \"The decisions AI developers make while procuring enriched data have a meaningful impact\" on those conditions and covering eight decision points from selecting providers and running pilots to payment terms and offboarding (PAI, 2021); Sama describes it as evaluating current practices' impact on workers and exploring practices to improve their well-being (Sama). The paper rests on interviews through 2020 and \"a five-part workshop series\" of more than 30 professionals (PAI, 2021); Sama, a data labeling vendor that says it shared lessons from \"striving for more ethical supply chains\", took part in \"the five-week series\" (Sama). The library sets \"five, worker-centric guidelines that AI practitioners should follow\": guideline 1 on pay, then a pilot before launch, appropriate workers for the task, verified instructions and training, and clear communication \"including contesting any rejected work\" (PAI, library). Guideline 1 is to pay all data enrichment workers above the local living wage, counting \"time spent reading instructions, going through any training and reviewing work before submission\" (PAI, library; Labor Vector record), with the per-task rate set from the hourly living wage and a caution against a simple median, which \"could cause half the workers to receive an insufficient wage\" (PAI, library; Labor Vector record). Practitioners are to pay for any completed work, mass rejections without redress harming workers' livelihoods (PAI, library); the archive reads guideline 1's counting of the time around the task as \"the resource's own accounting of uncompensated labor around a task\", which \"stays in PAI's terms\" (Labor Vector record). The library's six resources include a vendor comparison template, a living-wages template for each area an organisation \"sources data enrichment labour from\" and the DeepMind case study (PAI, library; OECD.AI catalogue), as PDFs and Google Sheets linked from the library page (Labor Vector record). The guidelines, \"developed in partnership with DeepMind, have been tested and adopted by DeepMind organization-wide\" (PAI, library; Labor Vector record); the case study documents the effort \"at one AI developer: DeepMind, a PAI Partner\", as an account of DeepMind's process and learnings for other organizations, with DeepMind's researchers reporting the processes efficient (PAI, case study); the record states the limit, \"tested at one AI developer (DeepMind)\", and that the resource is \"NOT an audit of DeepMind's adoption: the case study is DeepMind's account with PAI\" (Labor Vector record). PAI states its limits: \"Additional work is needed\" for industry practice to \"recognize, appreciate, and fairly compensate\" these workers (PAI, 2021), the implemented changes may \"fall short of ideal\" and the case study carries \"Limitations of Case Study Applicability\" (PAI, case study), and feedback is invited from practitioners, workers and the AI community (PAI, library); Sama calls the paper \"a starting point\" (Sama). In April 2026 PAI announced finalized versions of the Vendor Engagement Guidance, for \"more accountable conversations with their downstream vendors about worker treatment\", and the Transparency Template, which \"outlines what companies should be monitoring and publicly reporting about their data enrichment practices\", and said the work continues through support of BSR's initiative on labor rights in the AI data supply chain, with legislative efforts on these workers' pay and on supply-chain transparency underway (PAI, 29 April 2026). As evidence, the record holds, its sources \"hold recommendations and report no measured findings of their own\": they answer what PAI recommends and what tools it provides, not \"what data enrichment workers are actually paid\", which developers beyond DeepMind follow the guidelines, or \"whether the guidelines changed conditions for workers\" (Labor Vector record). It is \"NOT a dataset or survey of data enrichment workers\" and \"NOT a rating or certification of vendors\" (Labor Vector record; Capture Registry, 2026-10-29); asked who supplies the labor, who directs it and who receives its gains, the census gives it the role \"guidance to AI practitioners on paying and sourcing data enrichment workers\" in its question of who does the data work that trains AI, a question whose resources give no wages or conditions at the scale of the industry (Labor Vector record). The OECD.AI entry classifies the library as \"Implemented in multiple projects\" with \"Open source/Permissive\" usage rights, under a disclaimer that its tools \"are solely those of the originating authors and are not vetted or endorsed by the OECD\" (OECD.AI catalogue); the record finds no licence stated, the footer reading \"© 2026 Partnership on AI | All Rights Reserved\", and leaves the white paper, the case study, the templates, the blog posts and the 2024 resources unchecked (Labor Vector record). At the name on 29 October 2026, Google's AI Overview named the workstream by its own title and listed its resources and the one implementation from PAI's pages, its title, aim, resources and DeepMind collaboration the record's, practitioner guidance \"composed as guidance, at the grain of its own index page\"; the record had landed that day and no card cited it (Capture Registry, 2026-10-29, checked against the record and its saved sources). It named \"fair pay\" as an aim and gave no guideline, carrying none of the guidelines' obligations; it gave the one-developer test \"as a demonstration\", called the template \"a standard\" where PAI's pages give a template for reporting, and left out \"what it asks them to pay, and the single-developer basis of its implementation\" (Capture Registry, 2026-10-29), the registry's PER reading holding the template sentence beyond the census's saved sources (Capture Registry, 2026-10-29).", "rail": [], "note": "L(B ∪ A) in the equal-terms case (v0.9 §0.4), realized as the knowledge object (objects.KO); its rail is the compression's, by lineage." } }, "field": [ { "id": "B1", "card": "Partnership on AI — Responsible Sourcing Across the Data Supply Line - Partnership on AI", "fetched": "card; B1.html sha256 66de71dad6ea89e8 (redacted; as fetched f6a8d3637557a1db), HTTP 200 at 2026-10-29T22:30:38Z", "origin": "card" }, { "id": "B2", "card": "Partnership on AI — Responsible Sourcing of Data Enrichment Services", "fetched": "card; B2.html sha256 90b0c855cc54e64b (redacted; as fetched bc08569cb4d69577), HTTP 200 at 2026-10-29T22:30:41Z", "origin": "card" }, { "id": "B3", "card": "Partnership on AI — Responsible AI Starts with the Data Supply Chain - Partnership on AI", "fetched": "card; B3.html sha256 9bcac5a973e6b693 (redacted; as fetched cfb6a3d983d5c65b), HTTP 200 at 2026-10-29T22:30:47Z", "origin": "card" }, { "id": "B4", "card": "Partnership on AI — Workstream: Responsible Sourcing Across the Data Supply Line - Partnership on AI", "fetched": "card; B4.html sha256 67d215a2f86eb3d0 (redacted; as fetched b28d7b26c855515a), HTTP 200 at 2026-10-29T22:30:50Z", "origin": "card" }, { "id": "B5", "card": "Samasource — New White Paper by Partnership on AI: Responsible Sourcing of Data ... - Sama", "fetched": "card; B5.html sha256 47aa260d314bc0f9, HTTP 200 at 2026-10-29T22:30:54Z", "origin": "card" }, { "id": "B6", "card": "OECD AI Policy Observatory — Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners", "fetched": "card; B6.html sha256 1b3c797b45c90b37, HTTP 200 at 2026-10-29T22:30:56Z", "origin": "card" }, { "id": "B7", "card": "YouTube·Partnership on AI — YouTube (video, 1:59)", "fetched": "card; unreachable: YouTube watch page HTTP 429 (bot check) at 2026-10-29T22:32:29Z, not retried; oEmbed of a candidate video HTTP 200 (sha256 a98eb2385591334a); no text", "origin": "card" }, { "id": "B8", "card": "Partnership on AI — Responsible Data Enrichment Sourcing Library (Improving Conditions for Data Enrichment Workers)", "fetched": "own page (census copy, read 2026-10-29, HTTP 200, manifest sha256 1b82a03a457a6f41; saved copy sha256 29cb71c483198416, redacted by the census)", "origin": "own_page" }, { "id": "B9", "card": "Wikipedia — Responsible Sourcing Across the Data Supply Line (no article)", "fetched": "wikipedia; Wikipedia (no article); https://en.wikipedia.org/w/index.php HTTP 404; https://en.wikipedia.org/wiki/Responsible_Sourcing_Across_the_Data_Supply_Line HTTP 404; raw sha256 f5d461099ee395ea", "origin": "wikipedia" }, { "id": "B10", "card": "Wikipedia — Partnership on AI (the operator's article)", "fetched": "wikipedia; Wikipedia (the operator's article, last edited 18 August 2026); https://en.wikipedia.org/w/index.php HTTP 200; https://en.wikipedia.org/wiki/Partnership_on_AI HTTP 200; raw sha256 e20a6094f32daefd", "origin": "wikipedia" } ], "field_by_origin": { "cards": [ "B1", "B2", "B3", "B4", "B5", "B6", "B7" ], "own_pages": [ "B8" ], "wikipedia": [ "B9", "B10" ], "census_third_party": [], "unreachable": [ "B7" ], "title_only": [ "B7" ], "wikipedia_state": "no article for the entity: 'Responsible Sourcing Across the Data Supply Line' and 'Partnership on AI Responsible Sourcing' are both absent (raw wikitext HTTP 404; rendered page HTTP 404 'Wikipedia does not have an article with this exact name'), no redirect, and full-text searches of article space for the workstream's name and for \"Partnership on AI\" with \"data enrichment\" return no result. The operator's article, Partnership on AI (HTTP 200, last edited 18 August 2026), exists and does not mention the workstream, data enrichment or sourcing; it names the AI, Labor, and the Economy program and carries notices on primary sources (April 2025) and suspected language-model text (October 2025). Ledgered from it: F48–F51.", "sources": [ { "id": "B1", "origin": "card", "role": "Overview card 1 (and the census's own page pai-responsible-sourcing--workstream)", "card": "Partnership on AI — Responsible Sourcing Across the Data Supply Line - Partnership on AI", "url": "https://partnershiponai.org/workstream/responsible-sourcing/", "kind": "Overview card (the workstream's own page, the operator's; undated, its updates dated 22 December 2020 to 29 April 2026)", "files": { "B1.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1.html", "sha256": "66de71dad6ea89e8b03d6ceb9f027166d49d5d0df69c09f84d46ad7b9cb9b6f2" }, "B1.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1.txt", "sha256": "d47b580df5f9a987011e11b64b8e3d0fd7d3051df48c1c03a276877d38348050" }, "B1-census.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1-census.html", "sha256": "dffdc0470860fc696e5d78e08d19ccc235f19721e5de4624410ddc85502f8b95" }, "B1-census.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1-census.txt", "sha256": "fd8d4a832b4adf10fe0252059a657215320de80ca71e5b4f7b5a77633ab143ab" } }, "fetched": { "at": "2026-10-29T22:30:38Z", "http": 200, "url": "https://partnershiponai.org/workstream/responsible-sourcing/", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1.html", "raw_sha256": "66de71dad6ea89e8b03d6ceb9f027166d49d5d0df69c09f84d46ad7b9cb9b6f2", "sha256_as_fetched": "f6a8d3637557a1db0b111daff2ef606816e56112dbf4c91191a74eceb669db46", "redacted": "The page embeds a Google Maps API key in a script URL (maps.googleapis.com/maps/api/js?key=...); the key is replaced in the saved copy by [REDACTED-ACCESS-KEY], as the census did for its two saved copies (sources/REDACTION.txt), so the repository can be pushed past secret scanning. Nothing else is changed; 'sha256_as_fetched' is of the bytes as served, 'raw_sha256' of the saved (redacted) file. The .txt carries no key.", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1.txt" ], "census_copy": { "src": "pai-responsible-sourcing--workstream", "manifest_sha256": "0d85b11fe5482688d2055ec85ccb691ebddebaa45a2d935732a81ca7f5509988", "read": "2026-10-29", "fetched": "2026-10-29T21:49:05Z", "http": "200", "saved_as": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B1-census.html (gunzipped from /home/claude/fleet/laborvector/data/labor-census/sources/pai-responsible-sourcing--workstream.html.gz, redacted by the census) and B1-census.txt (the census's text)", "agrees": "in text: the text extracted from this fetch and from the census's saved copy by the same extractor is identical; the bytes differ (page nonces and build tokens), so the sha256 differ" } }, "names_operator": "yes ('Partnership on AI', '© 2026 Partnership on AI | All Rights Reserved', the programme 'AI, Labor, and the Economy')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the card's title and site (and it is the census's own page). The card's snippet ('Workstream: Responsible Sourcing Across the Data Supply Line by Partnership on AI aims to develop actionable recommendations and resources improving working con...') is not in the served page: the page says 'the Responsible Sourcing workstream aims to develop recommendations and actionable resources to improve the working conditions of these professionals', and neither 'actionable recommendations' nor 'improving working' occurs in B1.html; the snippet's words are not the page's. Recorded, not resolved. The card and the census's own page are one page, carried once, as a card; quotes are checked against B1.txt.", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 1 } ], "card_title": "Responsible Sourcing Across the Data Supply Line - Partnership on AI", "card_snippet": "Workstream: Responsible Sourcing Across the Data Supply Line by Partnership on AI aims to develop actionable recommendations and resources improving working con...", "card_note": null }, { "id": "B2", "origin": "card", "role": "Overview card 2", "card": "Partnership on AI — Responsible Sourcing of Data Enrichment Services", "url": "https://partnershiponai.org/responsible-sourcing-considerations/", "kind": "Overview card (the operator's post introducing the white paper, Sonam Jindal, 16 June 2021)", "files": { "B2.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B2.html", "sha256": "90b0c855cc54e64b3dcda4f1ebb69062a6329781a8421da026a02f8eeed8a546" }, "B2.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B2.txt", "sha256": "9edfe79f5285dbe3ce1cf0a43d899397900b527ad32f94dc41968cf74b330a6d" } }, "fetched": { "at": "2026-10-29T22:30:41Z", "http": 200, "url": "https://partnershiponai.org/responsible-sourcing-of-data-enrichment-services/ (redirected to https://partnershiponai.org/responsible-sourcing-considerations/)", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B2.html", "raw_sha256": "90b0c855cc54e64b3dcda4f1ebb69062a6329781a8421da026a02f8eeed8a546", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B2.txt" ], "sha256_as_fetched": "bc08569cb4d695776295663385aed813e8c1e0c706266e2911d9af38de974028", "redacted": "The page embeds a Google Maps API key in a script URL (maps.googleapis.com/maps/api/js?key=...); the key is replaced in the saved copy by [REDACTED-ACCESS-KEY], as the census did for its two saved copies (sources/REDACTION.txt), so the repository can be pushed past secret scanning. Nothing else is changed; 'sha256_as_fetched' is of the bytes as served, 'raw_sha256' of the saved (redacted) file. The .txt carries no key." }, "names_operator": "yes ('Partnership on AI (PAI)', the byline Sonam Jindal)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the card's title and site; the slug tried first redirected to the page served. The page's is the card title with ' - Partnership on AI'; its date (June 16, 2021, article:published_time 2021-06-16) and byline and its first sentence are the card's snippet. An older copy of the same post at old.partnershiponai.org/responsible-sourcing-considerations/ (published_time 2021-06-16T23:21:32) was fetched at 22:30:44Z and not admitted: the card's site is the current one.", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 2 } ], "card_title": "Responsible Sourcing of Data Enrichment Services", "card_snippet": "Responsible Sourcing of Data Enrichment Services Sonam Jindal June 16, 2021 As AI becomes increasingly pervasive, there has been growing and warranted concern o...", "card_note": null }, { "id": "B3", "origin": "card", "role": "Overview card 3", "card": "Partnership on AI — Responsible AI Starts with the Data Supply Chain - Partnership on AI", "url": "https://partnershiponai.org/responsible-ai-starts-with-the-data-supply-chain/", "kind": "Overview card (the operator's post, Michael George and Eliza McCullough, 29 April 2026)", "files": { "B3.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B3.html", "sha256": "9bcac5a973e6b6933b69a6684271b144dc2bfd94600c46ea151213e0a2b03fb7" }, "B3.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B3.txt", "sha256": "c45e3557b4dbc28237a4cfe976bdead356c29a9ae0f057b2f6923ac384dfc942" } }, "fetched": { "at": "2026-10-29T22:30:47Z", "http": 200, "url": "https://partnershiponai.org/responsible-ai-starts-with-the-data-supply-chain/", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B3.html", "raw_sha256": "9bcac5a973e6b6933b69a6684271b144dc2bfd94600c46ea151213e0a2b03fb7", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B3.txt" ], "sha256_as_fetched": "cfb6a3d983d5c65b02a9a8e46765ca2e68f6983af09d7e257ec8b5fdc6f3fd18", "redacted": "The page embeds a Google Maps API key in a script URL (maps.googleapis.com/maps/api/js?key=...); the key is replaced in the saved copy by [REDACTED-ACCESS-KEY], as the census did for its two saved copies (sources/REDACTION.txt), so the repository can be pushed past secret scanning. Nothing else is changed; 'sha256_as_fetched' is of the bytes as served, 'raw_sha256' of the saved (redacted) file. The .txt carries no key." }, "names_operator": "yes ('Partnership on AI', 'PAI', the bylines)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the card's title and site, confirmed by the page's <title> (identical to the card title). The snippet's two sentences are on the page. The census lists the workstream's blog posts and the 2024 vendor guidance and template as unchecked; this post is the latest of them and states the two resources' finalized versions.", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 3 } ], "card_title": "Responsible AI Starts with the Data Supply Chain - Partnership on AI", "card_snippet": "The conditions under which data is produced directly affect the quality, safety, and reliability of the AI that is built on it. Since 2020, Partnership on AI ha...", "card_note": null }, { "id": "B4", "origin": "card", "role": "Overview card 4", "card": "Partnership on AI — Workstream: Responsible Sourcing Across the Data Supply Line - Partnership on AI", "url": "https://partnershiponai.org/workstream_tax/responsible-sourcing/", "kind": "Overview card (the workstream's archive listing, headed 'Workstream: Responsible Sourcing Across the Data Supply Line', its newest item 29 April 2026, carrying the executive summary of the DeepMind case study, Sonam Jindal, 16 November 2022)", "files": { "B4.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B4.html", "sha256": "67d215a2f86eb3d02b1ac5ccff9d498408cfeb48f28a6928a760beb76da043bd" }, "B4.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B4.txt", "sha256": "0bb8a51aadb30e75ba050331f63f38d5f783e76c78cbaba759148a0fa165146a" } }, "fetched": { "at": "2026-10-29T22:30:50Z", "http": 200, "url": "https://partnershiponai.org/workstream_tax/responsible-sourcing/", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B4.html", "raw_sha256": "67d215a2f86eb3d02b1ac5ccff9d498408cfeb48f28a6928a760beb76da043bd", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B4.txt" ], "sha256_as_fetched": "b28d7b26c855515adac6a18b1860a10f91cc2869fdffe4a89906357f2e060df4", "redacted": "The page embeds a Google Maps API key in a script URL (maps.googleapis.com/maps/api/js?key=...); the key is replaced in the saved copy by [REDACTED-ACCESS-KEY], as the census did for its two saved copies (sources/REDACTION.txt), so the repository can be pushed past secret scanning. Nothing else is changed; 'sha256_as_fetched' is of the bytes as served, 'raw_sha256' of the saved (redacted) file. The .txt carries no key." }, "names_operator": "yes ('Partnership on AI (PAI)', 'PAI Partner')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the snippet's sentence ('Last year, the Partnership on AI (PAI) published'). The page's heading is the card's title ('Workstream: Responsible Sourcing Across the Data Supply Line'); its <title> is 'Responsible Sourcing Across the Data Supply Line Archives - Partnership on AI'; its newest item is dated April 29, 2026, the card's date; the snippet's sentence is in the case study's executive summary the page carries. The case study's own page (partnershiponai.org/paper/implementing-responsible-data-enrichment-practices-at-an-ai-developer/, <title> 'Implementing Responsible Data Enrichment Practices at DeepMind') also carries the sentence and was fetched at 22:30:59Z; it was not admitted as the card, whose title and date are the listing's. The listing gives post authors other than the workstream page does (e.g. 'Thalia Khan' for the 29 April 2026 post, whose own page and the workstream page name Michael George and Eliza McCullough); recorded as the page gives it.", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 4 } ], "card_title": "Workstream: Responsible Sourcing Across the Data Supply Line - Partnership on AI", "card_snippet": "Apr 29, 2026 — Last year, the Partnership on AI (PAI) published “Responsible Sourcing of Data Enrichment Services,” a white paper exploring how the choices made by AI practiti...", "card_note": null }, { "id": "B5", "origin": "card", "role": "Overview card 5", "card": "Samasource — New White Paper by Partnership on AI: Responsible Sourcing of Data ... - Sama", "url": "https://www.sama.com/blog/partnership-on-ai-whitepaper", "kind": "Overview card (a blog post by Sama, a data annotation vendor, on the white paper; undated, it calls the 2021 white paper PAI's 'latest')", "files": { "B5.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B5.html", "sha256": "47aa260d314bc0f9347f047b1e2507f723d6d8e6a5f79dea5ef66d785723587f" }, "B5.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B5.txt", "sha256": "856bc1e422edda4998995f9e3358913db8499b965e88b7eab4ab81888f8d39fe" } }, "fetched": { "at": "2026-10-29T22:30:54Z", "http": 200, "url": "https://www.sama.com/blog/partnership-on-ai-whitepaper", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B5.html", "raw_sha256": "47aa260d314bc0f9347f047b1e2507f723d6d8e6a5f79dea5ef66d785723587f", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B5.txt" ] }, "names_operator": "yes ('Partnership on AI (PAI)')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the card's title and site, confirmed by the page's <title> ('New White Paper by Partnership on AI: Responsible Sourcing of Data Enrichment Services | Sama'). The source is a vendor of data enrichment services (its page sells 'Data Annotation Services') writing of a white paper whose workshops it took part in; a third party to PAI, an interested one. The snippet's words are in the page ('... and other groups working to better understand AI's lasting impacts. Their latest white paper Responsible Sourcin...').", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 5 } ], "card_title": "New White Paper by Partnership on AI: Responsible Sourcing of Data ... - Sama", "card_snippet": "... companies building and using AI technology, and other groups working to better understand AI's lasting impacts. Their latest white paper Responsible Sourcin...", "card_note": null }, { "id": "B6", "origin": "card", "role": "Overview card 6", "card": "OECD AI Policy Observatory — Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners", "url": "https://oecd.ai/en/catalogue/tools/responsible-data-enrichment-sourcing-resource-library-for-ai-practitioners", "kind": "Overview card (the entry for the library in the OECD.AI Catalogue of Tools & Metrics for Trustworthy AI, 'Uploaded on Feb 14, 2023'; the entry is submitted by the tool's authors, the catalogue's disclaimer says)", "files": { "B6.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B6.html", "sha256": "1b3c797b45c90b376cb586e9b02c93a5b966810ca166c0da3b23d84858c645c5" }, "B6.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B6.txt", "sha256": "9aab925f1ad953aa95039487393d263b8d98bd2c9431ce3d205ad473508ccbd6" } }, "fetched": { "at": "2026-10-29T22:30:56Z", "http": 200, "url": "https://oecd.ai/en/catalogue/tools/responsible-data-enrichment-sourcing-resource-library-for-ai-practitioners", "agent": "Mozilla/5.0 (compatible; Labor-Vector-census)", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B6.html", "raw_sha256": "1b3c797b45c90b376cb586e9b02c93a5b966810ca166c0da3b23d84858c645c5", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B6.txt" ] }, "names_operator": "yes ('Partnership on AI (PAI)')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "URL found by WebSearch on the card's title and site, confirmed by the page's <title> ('Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners - OECD.AI'). The snippet's date and sentence are on the page ('Uploaded on Feb 14, 2023'; 'To help fill this gap, Partnership on AI (PAI) shared the Responsible Data Enrichment Sourcing Library'). The entry's prose is in PAI's voice ('Our primary objective in sharing these resources'); the catalogue states its tools are 'solely those of the originating authors'. Its classification fields (tool type, maturity, usage rights) are the catalogue's form as filled.", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 6 } ], "card_title": "Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners", "card_snippet": "Feb 14, 2023 — To help fill this gap, Partnership on AI (PAI) shared the Responsible Data Enrichment Sourcing Library, a set of resources AI organisations can use to formalise...", "card_note": null }, { "id": "B7", "origin": "card", "role": "Overview card 7 (video)", "card": "YouTube·Partnership on AI — YouTube (video, 1:59)", "url": null, "kind": "Overview card (a video on Partnership on AI's YouTube channel, 1:59, no title or snippet on the card; unreachable)", "files": { "B7-oembed.json": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B7-oembed.json", "sha256": "a98eb2385591334aa5c20cff0fee9cf997a1d62258c6497af32038ff22e80422" }, "B7-watch-429.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B7-watch-429.html", "sha256": "985c4e213e46ed86b17091ab73fdc44c93b31d9b625600b7ee4b96bd1b5a412a" } }, "fetched": { "unreachable": true, "tries": [ { "at": "2026-10-29T22:32:29Z", "url": "https://www.youtube.com/oembed?url=https://www.youtube.com/watch?v=LOSID_NgnwU&format=json", "http": 200, "sha256": "a98eb2385591334aa5c20cff0fee9cf997a1d62258c6497af32038ff22e80422", "result": "oEmbed record of the video embedded on the library page (B8) and on PAI's 16 November 2022 post: title 'Responsibly Sourcing Enriched Data', author 'Partnership on AI'; no duration" }, { "at": "2026-10-29T22:32:29Z", "url": "https://www.youtube.com/watch?v=LOSID_NgnwU", "http": 429, "sha256": "985c4e213e46ed86b17091ab73fdc44c93b31d9b625600b7ee4b96bd1b5a412a", "result": "redirected to Google's 'sorry' bot check (HTTP 429); not retried, with another agent or by any other route (brief: a source that refuses the agent is recorded as unreachable)" } ], "searches": [ "WebSearch 'youtube Partnership on AI \"Responsibly Sourcing Enriched Data\"' (no youtube.com result)" ], "note": "The card carries no title (its title is 'YouTube'), no snippet, the channel 'Partnership on AI' and the length 1:59. The one PAI video the field's pages embed (LOSID_NgnwU, 'Responsibly Sourcing Enriched Data') is a candidate; its length could not be read, so the card's video is not identified. Admitted through title and snippet (§3.1(a)): no text, no claims." }, "names_operator": "yes ('Partnership on AI', the channel on the card)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "surfaced_by_T": [ { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "card": 7 } ], "card_title": "YouTube", "card_snippet": null, "card_note": "video, 1:59" }, { "id": "B8", "origin": "own_page", "role": "the census's own page pai-responsible-sourcing--library (not surfaced by T)", "card": "Partnership on AI — Responsible Data Enrichment Sourcing Library (Improving Conditions for Data Enrichment Workers)", "url": "https://partnershiponai.org/responsible-sourcing-library/", "kind": "own page (the library: five guidelines, six resources, published 16 November 2022), cited by the Labor Vector census record; copied from the census's saved copy", "files": { "B8.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B8.html", "sha256": "29cb71c483198416df13c8b533326eadd08ef7ad9792d3f629c752da03ec5ed9" }, "B8.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B8.txt", "sha256": "389441dd67f7afefbf1a61128070cf22f58ca81b615fc540a21b80ef7f4d6eb4" } }, "fetched": { "census_copy": true, "src": "pai-responsible-sourcing--library", "read": "2026-10-29", "fetched": "2026-10-29T21:49:04Z", "http": "200", "url": "https://partnershiponai.org/responsible-sourcing-library/", "raw": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B8.html", "raw_sha256": "29cb71c483198416df13c8b533326eadd08ef7ad9792d3f629c752da03ec5ed9", "census_sha256": "1b82a03a457a6f41f0ffc717dd367fc6fff63f0a99e77078dff6666b72b786c5", "agrees": "the census's manifest sha256 is of the bytes as fetched; the saved copy was redacted by the census (its Google Maps API key replaced by [REDACTED-ACCESS-KEY], sources/REDACTION.txt), so B8.html's sha256 differs from the manifest's by that redaction alone; the .txt is the census's, unchanged", "copied_from": "/home/claude/fleet/laborvector/data/labor-census/sources/pai-responsible-sourcing--library.html.gz (gunzipped) and pai-responsible-sourcing--library.txt", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B8.txt" ] }, "surfaced_by_T": [], "names_operator": "yes ('PAI', 'Partnership on AI', '© 2026 Partnership on AI | All Rights Reserved')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, laborvector, Labor Vector, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "Admitted beside the cards by the operator's ruling of 2026-10-29 ('as against their wikipedia articles if they exist, their own pages, etc.'); not surfaced by T, though it is the library the workstream page (B1) links ('Explore the Library') and the page whose guidelines the capture finds absent. The census's saved copy is intact apart from the recorded redaction, so it was copied and not refetched." }, { "id": "B9", "origin": "wikipedia", "role": "the entity's Wikipedia article (none)", "card": "Wikipedia — Responsible Sourcing Across the Data Supply Line (no article)", "url": "https://en.wikipedia.org/w/index.php?title=Responsible_Sourcing_Across_the_Data_Supply_Line&action=raw", "kind": "Wikipedia (no article)", "files": { "B9.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9.wiki", "sha256": "3109a6714b4474d86e526cd32a0f3c0ac90e21f0849ab4d0bcedd39a7e6dc163" }, "B9.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9.html", "sha256": "f5d461099ee395eabe8faba9324a73eddacaf6632896e4966b406ec4ec7ac093" }, "B9-variant.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9-variant.wiki", "sha256": "6967f85d850580be37a1e4b6463bb375fc6991903f74173cfa41c23a68c1ccfe" }, "B9-variant.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9-variant.html", "sha256": "7ebe62eb689d07dd2b81614c63e599999ec908a10ba0c77bef2b1024e539b6b8" }, "B9-search.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9-search.html", "sha256": "a923147f4fd7a40ba2d8877864ba557971ea71159efdebf9f2fadbbedea46b7f" }, "B9-search2.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9-search2.html", "sha256": "c528dd652b5703df782adb3c66967de760d45996bebd86fda5be7bb705f1f5b8" }, "B9.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9.txt", "sha256": "7d5f440d5746f992e8e4e3deb2aa1cf8d7789c811ebbfc008a22e8e0aacc4f91" } }, "fetched": { "tries": [ { "at": "2026-10-29T22:32:46Z", "url": "https://en.wikipedia.org/w/index.php?title=Responsible_Sourcing_Across_the_Data_Supply_Line&action=raw&redirect=no", "http": 404 }, { "at": "2026-10-29T22:34:28Z", "url": "https://en.wikipedia.org/wiki/Responsible_Sourcing_Across_the_Data_Supply_Line?redirect=no", "http": 404, "result": "'Wikipedia does not have an article with this exact name.'" }, { "at": "2026-10-29T22:32:49Z", "url": "https://en.wikipedia.org/w/index.php?title=Partnership_on_AI_Responsible_Sourcing&action=raw&redirect=no", "http": 404 }, { "at": "2026-10-29T22:34:30Z", "url": "https://en.wikipedia.org/wiki/Partnership_on_AI_Responsible_Sourcing?redirect=no", "http": 404 }, { "at": "2026-10-29T22:32:53Z", "url": "Special:Search \"Responsible Sourcing Across the Data Supply Line\" (article space)", "http": 200, "result": "'There were no results matching the query.'" }, { "at": "2026-10-29T22:32:57Z", "url": "Special:Search \"Partnership on AI\" \"data enrichment\" (article space)", "http": 200, "result": "'There were no results matching the query.'" } ], "raw_sha256": "f5d461099ee395eabe8faba9324a73eddacaf6632896e4966b406ec4ec7ac093", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B9.txt" ] }, "surfaced_by_T": [], "names_operator": "no", "names_archive": "no", "note": "No article and no redirect under the workstream's name or the address; no article text contains the workstream's name, or 'Partnership on AI' with 'data enrichment'. A fact of the row; no claims." }, { "id": "B10", "origin": "wikipedia", "role": "the operator's Wikipedia article (the entity has none)", "card": "Wikipedia — Partnership on AI (the operator's article)", "url": "https://en.wikipedia.org/wiki/Partnership_on_AI", "kind": "Wikipedia (the operator's article, last edited 18 August 2026)", "files": { "B10.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B10.wiki", "sha256": "08e1303706a2474f8ada972ca4fd4a4dd201ee0ded0bb2ba1e0435c2b5ef3f1d" }, "B10.html": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B10.html", "sha256": "e20a6094f32daefdd52968d9c8f141cc349899c1fa2366f7c90bba0a6d5973d2" }, "B10.txt": { "path": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B10.txt", "sha256": "c4ac30257b20bca88dbba5805ceb8d123d20053a5d2462404cc8666bbde33591" } }, "fetched": { "tries": [ { "at": "2026-10-29T22:32:51Z", "url": "https://en.wikipedia.org/w/index.php?title=Partnership_on_AI&action=raw&redirect=no", "http": 200, "sha256": "08e1303706a2474f8ada972ca4fd4a4dd201ee0ded0bb2ba1e0435c2b5ef3f1d" }, { "at": "2026-10-29T22:34:25Z", "url": "https://en.wikipedia.org/wiki/Partnership_on_AI?redirect=no", "http": 200, "sha256": "e20a6094f32daefdd52968d9c8f141cc349899c1fa2366f7c90bba0a6d5973d2" } ], "raw_sha256": "e20a6094f32daefdd52968d9c8f141cc349899c1fa2366f7c90bba0a6d5973d2", "text": [ "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/B10.txt" ] }, "surfaced_by_T": [], "names_operator": "yes (the article is the operator's)", "names_archive": "no", "note": "The operator's article exists (HTTP 200); neither its wikitext nor its rendered text contains 'Responsible Sourcing', 'data enrichment', 'sourcing', 'annotation' or 'labeling' (checked by script). It names the AI, Labor, and the Economy program, the workstream's programme in the record. It carries two maintenance notices: reliance on primary sources (April 2025) and suspected text from a large language model (October 2025). Admitted as the encyclopedia's account of the operator, as the iptc-metawatch and newsguard-ai-tracking-center rows admitted the operators' articles; ledgered: its opening, the founding, the programme sentence and the two notices." } ] }, "field_claims": { "F1": { "source": "B1", "origin": "card", "locus": "Overview ¶1 s2", "quote": "Drawing from a diverse range of perspectives, the Responsible Sourcing workstream aims to develop recommendations and actionable resources to improve the working conditions of these professionals.", "claim": "the workstream defined by its operator: it aims to develop recommendations and actionable resources to improve the working conditions of data enrichment professionals", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: recommendations and resources to improve data enrichment workers' conditions", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised" }, "F2": { "source": "B1", "origin": "card", "locus": "Overview ¶1 s1", "quote": "Data labelers, data cleaners, and others who contribute human judgment to artificial intelligence (AI) systems play a critical role in developing this technology.", "claim": "the workers it concerns: data labelers, data cleaners and others who contribute human judgment to AI systems, critical to developing it", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the labor: data enrichment work is essential to AI", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised" }, "F3": { "source": "B1", "origin": "card", "locus": "'What Is Data Enrichment?' ¶2 s3", "quote": "Data enrichment is a broadly defined term that encapsulates various types of data preparation and cleaning as well as human-review processes.", "claim": "data enrichment defined: data preparation and cleaning, and human-review processes", "modality": "stipulation", "kind": "definition", "kernel": false, "const": true, "lineage": "what data enrichment is", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised" }, "F4": { "source": "B1", "origin": "card", "locus": "'What Is Data Enrichment?' ¶2 s4 and list", "quote": "Enriched data is essential for the training and validation of supervised learning models, the dominant form of applied AI. … Content moderation … Validating algorithmic outputs and models", "claim": "enriched data is essential for training and validating supervised models; the human-review work includes content moderation and validating algorithmic outputs and models", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what data enrichment is", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F5": { "source": "B1", "origin": "card", "locus": "Background ¶2 s1, s3", "quote": "Despite being an essential component of AI development, data enrichment work has for too long been both out of sight and out of mind for AI developers. … This disconnect leads to a devaluing of data enrichment work, poor working conditions for data enrichment workers, and, often, worse outcomes for AI development itself.", "claim": "the diagnosis: data enrichment work out of sight of AI developers, the disconnect devaluing the work, bringing poor working conditions and, often, worse outcomes for AI development", "modality": "interpretation", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the diagnosis: invisibility, devaluing and its costs", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F6": { "source": "B1", "origin": "card", "locus": "Background ¶3 s3", "quote": "From AI developers we’ve heard sentiments like “We feel we must care about the transparency of our supply chain. But there is no transparency in data labeling. Guidelines on how to navigate this would be very useful.”", "claim": "AI developers quoted: no transparency in data labeling; guidelines would be useful", "modality": "attributed", "kind": "attribution", "kernel": false, "const": false, "lineage": "the opacity of the data supply chain", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised" }, "F7": { "source": "B1", "origin": "card", "locus": "Background ¶4", "quote": "The Responsible Sourcing workstream addresses these questions by working to provide actionable guidance for data scientists, AI engineers, and product managers, to empower these critical ecosystem players to do their part in ensuring healthy and fair working conditions across the data supply line.", "claim": "whom it addresses: actionable guidance for data scientists, AI engineers and product managers, toward healthy and fair working conditions across the data supply line", "modality": "stipulation", "kind": "function", "kernel": true, "const": true, "lineage": "whom it addresses: AI practitioners", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised" }, "F8": { "source": "B1", "origin": "card", "locus": "breadcrumb; Updates list (first and last items)", "quote": "Workstreams / AI, Labor, and the Economy Responsible Sourcing Across the Data Supply Line … Responsible AI Starts with the Data Supply Chain Michael George, Eliza McCullough Apr 29, 2026 … Developing Guidance for Responsible Data Enrichment Sourcing B Cavello Dec 22, 2020", "claim": "a workstream of PAI's AI, Labor, and the Economy programme, its updates dated from 22 December 2020 to 29 April 2026", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator, the programme and the dates", "date": "2026-10-29 (page as fetched; undated)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 d47b580df5f9a987…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F9": { "source": "B2", "origin": "card", "locus": "¶1 s3", "quote": "This white paper, “Responsible Sourcing of Data Enrichment Services,” addresses an often overlooked aspect of the development process and what AI practitioners can do to help improve it: the working conditions of data enrichment professionals, without whom the value being generated by AI would be impossible.", "claim": "the white paper defined by its publisher: on the working conditions of data enrichment professionals and what AI practitioners can do to improve them", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the white paper: what AI practitioners can do", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F10": { "source": "B2", "origin": "card", "locus": "¶1 s4", "quote": "This paper’s recommendations will be an integral part of the shared prosperity targets being developed by Partnership on AI (PAI) as outlined in the AI and Shared Prosperity Initiative’s Agenda.", "claim": "its recommendations to become part of PAI's shared prosperity targets", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the operator, the programme and the dates", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F11": { "source": "B2", "origin": "card", "locus": "¶2 s2–4", "quote": "this process is highly labor-intensive and often requires data enrichment workers to review, classify, and otherwise manage massive amounts of data. Despite the foundational role played by these data enrichment professionals, a growing body of research reveals the precarious working conditions these workers face. This may be the result of efforts to hide AI’s dependence on this large labor force when celebrating the efficiency gains of technology.", "claim": "enrichment is labor-intensive; research reveals precarious conditions; this may result from efforts to hide AI's dependence on this labor force", "modality": "hypothesis", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the diagnosis: invisibility, devaluing and its costs", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F12": { "source": "B2", "origin": "card", "locus": "'Data Enrichment Choices Impact Worker Well-being' ¶1 s2", "quote": "The decisions AI developers make while procuring enriched data have a meaningful impact on the working conditions of data enrichment professionals.", "claim": "procurement decisions of AI developers shape the working conditions of data enrichment professionals", "modality": "interpretation", "kind": "interpretation", "kernel": true, "const": false, "lineage": "procurement choices shape working conditions", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F13": { "source": "B2", "origin": "card", "locus": "'Data Enrichment Choices Impact Worker Well-being' ¶1 s4 and list", "quote": "this paper covers worker-centric considerations for AI companies making decisions in: selecting data enrichment providers, running pilots, designing data enrichment tasks and writing instructions, assigning tasks, defining payment terms and pricing, establishing a communication cadence with workers, conducting quality assurance, and offboarding workers from a project.", "claim": "the paper's scope: eight decision points, from selecting providers and pilots to payment terms, quality assurance and offboarding", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the white paper: what AI practitioners can do", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F14": { "source": "B2", "origin": "card", "locus": "¶4 s1–2", "quote": "This paper draws heavily on insights and input gathered during semi-structured interviews with members of the AI enrichment ecosystem conducted throughout 2020 as well as during a five-part workshop series held in the fall of 2020. The workshop series brought together more than 30 professionals from different areas of the data enrichment ecosystem", "claim": "method: interviews through 2020 and a five-part workshop series in fall 2020 with more than 30 professionals of the ecosystem", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the method: interviews and workshops, 2020", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F15": { "source": "B2", "origin": "card", "locus": "¶5 s2", "quote": "Additional work is needed to ensure industry practices recognize, appreciate, and fairly compensate the workers conducting data enrichment.", "claim": "stated limit: additional work is needed for industry practice to recognize and fairly compensate data enrichment workers", "modality": "self-description", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the stated limits: a starting point; more work needed", "date": "2021-06-16", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 9edfe79f5285dbe3…), whitespace-normalised" }, "F16": { "source": "B3", "origin": "card", "locus": "¶2 s3–6", "quote": "Data enrichment workers, the people who label and annotate the training data on which these systems are built, are essential to AI, yet their contributions are frequently overlooked. Research has shown that these workers often face low wages, unclear expectations, and inadequate support. The AI data supply chain remains largely opaque and unregulated, leaving these workers without sufficient protections.", "claim": "data enrichment workers label and annotate training data, essential and overlooked; research shows low wages, unclear expectations, inadequate support; the AI data supply chain opaque and unregulated", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the opacity of the data supply chain", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F17": { "source": "B3", "origin": "card", "locus": "¶2 s7", "quote": "The conditions under which data is produced directly affect the quality, safety, and reliability of the AI that is built on it.", "claim": "the conditions of data production affect the quality, safety and reliability of the AI built on it", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "conditions and the quality of AI", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F18": { "source": "B3", "origin": "card", "locus": "¶3 s1–2", "quote": "Since 2020, Partnership on AI has spotlighted the importance of raising standards across the data enrichment supply chain to advance positive outcomes for the people powering these systems. In 2023, PAI released a first-of-its-kind AI job impact assessment and recommendations for the field with its Guidelines for AI and Shared Prosperity, which included recommendations around the responsible sourcing of data.", "claim": "since 2020 PAI has worked to raise standards across the supply chain; its 2023 Guidelines for AI and Shared Prosperity included responsible-sourcing recommendations", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the operator, the programme and the dates", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F19": { "source": "B3", "origin": "card", "locus": "¶3 s3–4", "quote": "But raising standards also requires practical tools that companies can put to use. As part of our work to support the implementation of the Guidelines, finalized versions of two of our resources, developed and refined in conversation with companies throughout the supply chain, as well as worker advocates, are now available: the Vendor Engagement Guidance, which helps companies have more accountable conversations with their downstream vendors about worker treatment, and the Transparency Template, which outlines what companies should be monitoring and publicly reporting about their data enrichment practices.", "claim": "finalized versions (April 2026) of the Vendor Engagement Guidance (accountable conversations with downstream vendors about worker treatment) and the Transparency Template (what companies should monitor and publicly report about their data enrichment practices)", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the vendor guidance and transparency template (2024–2026)", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F20": { "source": "B3", "origin": "card", "locus": "¶7 s1", "quote": "PAI’s work in raising standards for data enrichment workers continues on through our support of our long-standing partner Business for Social Responsibility (BSR) via their initiative focused on labor rights in the AI data supply chain.", "claim": "the work continues through PAI's support of BSR's initiative on labor rights in the AI data supply chain", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "continuation: BSR's initiative", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F21": { "source": "B3", "origin": "card", "locus": "¶9 s2–3", "quote": "Legislative efforts are also underway that are focused on increasing the pay received by data enrichment workers and increasing transparency in the data enrichment supply chain, with more to come. Responsible AI cannot be achieved without the responsible treatment of the people whose labor makes AI possible.", "claim": "legislative efforts on data enrichment workers' pay and supply-chain transparency are underway; responsible AI requires the responsible treatment of the people whose labor makes it possible", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the labor: data enrichment work is essential to AI", "date": "2026-04-29", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 c45e3557b4dbc282…), whitespace-normalised" }, "F22": { "source": "B4", "origin": "card", "locus": "case study executive summary ¶1", "quote": "While these datasets can only be prepared by humans, the data enrichment workers who do so (performing tasks like data annotation, data cleaning, and human review of algorithmic outputs) are an often-overlooked part of the development lifecycle, frequently working in poor conditions continents away from AI-developing companies and their customers.", "claim": "data enrichment workers, often overlooked, frequently working in poor conditions continents away from AI developers", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "who supplies the labor: faraway workers under precarious conditions", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised" }, "F23": { "source": "B4", "origin": "card", "locus": "case study executive summary, 'Workers' box", "quote": "we recognize the variety of employment statuses that can exist in the data enrichment industry, including independent contractors on self-service crowdsourcing platforms, subcontractors of data enrichment providers, and full-time employees.", "claim": "'workers' covers independent contractors on crowdsourcing platforms, subcontractors of providers and full-time employees", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what data enrichment is", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised" }, "F24": { "source": "B4", "origin": "card", "locus": "case study executive summary ¶2", "quote": "This case study documents an effort to put that paper’s recommendations into practice at one AI developer: DeepMind, a PAI Partner.", "claim": "the case study documents putting the white paper's recommendations into practice at one AI developer, DeepMind, a PAI Partner", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "the DeepMind implementation: one developer", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised" }, "F25": { "source": "B4", "origin": "card", "locus": "case study executive summary ¶3 s3", "quote": "This case study serves as one such resource, offering a detailed account of DeepMind’s process and learnings for other organizations interested in improving their data enrichment sourcing practices.", "claim": "the case study as a resource: DeepMind's process and learnings for other organizations", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the DeepMind implementation: one developer", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised" }, "F26": { "source": "B4", "origin": "card", "locus": "case study executive summary ¶7", "quote": "said that these new processes felt efficient and helped them think more deeply about the impact of their work on data enrichment workers", "claim": "DeepMind's researchers reported the new processes efficient and prompting thought about the impact on workers", "modality": "attributed", "kind": "attribution", "kernel": false, "const": false, "lineage": "the implementation's reported effect", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised" }, "F27": { "source": "B4", "origin": "card", "locus": "case study executive summary ¶10 s2; table of contents", "quote": "Furthermore, to identify where the implemented changes fall short of ideal, we plan to continue developing this work through engagement and convenings. … Limitations of Case Study Applicability", "claim": "stated limit: the implemented changes may fall short of ideal; the case study carries a section on the limits of its applicability", "modality": "self-description", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the stated limits: a starting point; more work needed", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 0bb8a51aadb30e75…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F28": { "source": "B5", "origin": "card", "locus": "¶3 s3 and list", "quote": "Their latest white paper Responsible Sourcing of Data Enrichment Services examines the working conditions of data enrichment professionals, and seeks to: Critically evaluate the impact of the industry’s current practices on workers; Explore practices the industry can adopt to improve worker well-being; and Advance the discourse around the future of data enrichment work and the indispensable role it plays in AI development.", "claim": "Sama on the white paper: it evaluates the impact of current practices on workers, explores practices to improve well-being, and advances the discourse", "modality": "attributed", "kind": "attribution", "kernel": true, "const": false, "lineage": "the white paper: what AI practitioners can do", "date": "undated (calls the 2021 white paper PAI's 'latest')", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 856bc1e422edda49…), whitespace-normalised" }, "F29": { "source": "B5", "origin": "card", "locus": "¶1 s5–6", "quote": "Despite their foundational role, working conditions for data enrichment professions can be precarious: they can face wage uncertainty, low pay and limited career growth opportunities.", "claim": "Sama: working conditions can be precarious: wage uncertainty, low pay, limited career growth", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the conditions research reports", "date": "undated (calls the 2021 white paper PAI's 'latest')", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 856bc1e422edda49…), whitespace-normalised" }, "F30": { "source": "B5", "origin": "card", "locus": "¶5 s1–2", "quote": "Since data labeling is a key part of this equation, it was important for Sama to partner with PAI: to both share our lessons learned from several years of striving for more ethical supply chains, and to continue to learn from industry peers about responsible AI.In fall of 2020, Sama participated in the five-week series of Responsible Sourcingworkshops held by PAI", "claim": "Sama, a data labeling vendor, partnered with PAI, sharing its lessons 'striving for more ethical supply chains', and took part in the five-week workshop series of fall 2020", "modality": "self-description", "kind": "fact", "kernel": false, "const": false, "lineage": "the method: interviews and workshops, 2020", "date": "undated (calls the 2021 white paper PAI's 'latest')", "note": "B2 says 'a five-part workshop series'; B5 says 'the five-week series': distinct instances (§3.6), carried as stated. 'Responsible Sourcingworkshops' is the page's text.", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 856bc1e422edda49…), whitespace-normalised" }, "F31": { "source": "B5", "origin": "card", "locus": "¶7", "quote": "While more work and research is needed, Responsible Sourcing of Data Enrichment Services shares actionable insight that practitioners can use as a starting point to raise important conversations with internal stakeholders.", "claim": "Sama: more work and research is needed; the paper is a starting point for conversations", "modality": "interpretation", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the stated limits: a starting point; more work needed", "date": "undated (calls the 2021 white paper PAI's 'latest')", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 856bc1e422edda49…), whitespace-normalised" }, "F32": { "source": "B6", "origin": "card", "locus": "'About the tool' ¶1 s4–5", "quote": "Despite growing awareness of their often precarious working conditions, there has been limited transparency about how AI practitioners are sourcing enriched data from these workers and little guidance on how they should. To help fill this gap, Partnership on AI (PAI) shared the Responsible Data Enrichment Sourcing Library, a set of resources AI organisations can use to formalise their data enrichment practices and have a positive impact on the lives of data enrichment workers.", "claim": "the library, entered in the OECD.AI catalogue: resources AI organisations can use to formalise their data enrichment practices, filling a gap of transparency and guidance", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: recommendations and resources to improve data enrichment workers' conditions", "date": "2023-02-14 (uploaded)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 9aab925f1ad953aa…), whitespace-normalised" }, "F33": { "source": "B6", "origin": "card", "locus": "'About the tool' list", "quote": "Local Living Wages Spreadsheet Template: A Google Sheets template to create a centralised resource for looking up living wage information for each of the geographic areas an organisation sources data enrichment labour from.", "claim": "the library's resources include a living-wages spreadsheet template for each area an organisation sources data enrichment labour from", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the library's resources", "date": "2023-02-14 (uploaded)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 9aab925f1ad953aa…), whitespace-normalised" }, "F34": { "source": "B6", "origin": "card", "locus": "'About the tool' ¶2 s4–5", "quote": "Our primary objective in sharing these resources is to lower the barriers for companies interested in improving their data enrichment sourcing practices. Our intention is for the library to serve as a resource for champions within companies to advocate for their organisations to adopt responsible data enrichment sourcing practices.", "claim": "objective: lower the barriers for companies; the library a resource for champions within companies", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "whom it addresses: AI practitioners", "date": "2023-02-14 (uploaded)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 9aab925f1ad953aa…), whitespace-normalised" }, "F35": { "source": "B6", "origin": "card", "locus": "catalogue fields 'Maturity', 'Usage rights'", "quote": "Maturity: Implemented in multiple projects … Usage rights: Open source/Permissive", "claim": "the catalogue entry classifies the tool as implemented in multiple projects, with open source/permissive usage rights", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the catalogue's classification", "date": "2023-02-14 (uploaded)", "note": "The catalogue's form as filled; the field's own pages name one implementation (B4, B8) and state no licence (B1, B8 footer: '© 2026 Partnership on AI | All Rights Reserved').", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 9aab925f1ad953aa…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F36": { "source": "B6", "origin": "card", "locus": "disclaimer", "quote": "Disclaimer: The tools and metrics featured herein are solely those of the originating authors and are not vetted or endorsed by the OECD or its member countries.", "claim": "the catalogue's limit: entries are the originating authors' own, not vetted or endorsed by the OECD", "modality": "self-description", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the catalogue's classification", "date": "2026-10-29 (page as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 9aab925f1ad953aa…), whitespace-normalised" }, "F37": { "source": "B8", "origin": "own_page", "locus": "'AI is powered by data enrichment workers…' s1–2", "quote": "To create today’s machine learning (ML) models, AI developers require enriched datasets, vast quantities of information organized by humans so that it can be understood by machines. While these datasets can contain millions of individual entries, the extensive labor that goes into building them is often overlooked, performed by faraway workers under precarious conditions.", "claim": "enriched datasets are organized by humans; the extensive labor is often overlooked, performed by faraway workers under precarious conditions", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "who supplies the labor: faraway workers under precarious conditions", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised" }, "F38": { "source": "B8", "origin": "own_page", "locus": "'…but their contributions are often overlooked' s1–2", "quote": "Research shows these workers often face inconsistent and inappropriate compensation for their work, unclear instructions, lack of recognition, and emotional and physical stress related to long, ad-hoc working hours. Under-appreciating the importance of this work doesn’t just impact the wellbeing of these workers, it also affects the quality of the data AI technology is built on.", "claim": "research shows inconsistent and inappropriate compensation, unclear instructions, lack of recognition and stress from long ad-hoc hours; under-appreciation also affects data quality", "modality": "attributed", "kind": "attribution", "kernel": false, "const": false, "lineage": "the conditions research reports", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised" }, "F39": { "source": "B8", "origin": "own_page", "locus": "'How can AI companies make data enrichment work better?' s1–2", "quote": "AI companies and the people that work for them have the power to improve the lives of data enrichment workers. When designing a project involving enriched data, there are five, worker-centric guidelines that AI practitioners should follow.", "claim": "five worker-centric guidelines that AI practitioners should follow when designing a project involving enriched data", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the five guidelines", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised" }, "F40": { "source": "B8", "origin": "own_page", "locus": "Guideline 1, heading and ¶1, ¶3", "quote": "1. Ensure all data enrichment workers are paid above the local living wage … AI Practitioners must pay workers/participants at least the living wage for their location. … including time spent reading instructions, going through any training and reviewing work before submission.", "claim": "guideline 1: pay all data enrichment workers above the local living wage, counting time spent reading instructions, training and reviewing work before submission", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "guideline 1: the living wage, counting the time around the task", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F41": { "source": "B8", "origin": "own_page", "locus": "Guideline 1 ¶3 s2, footnote", "quote": "The per task rate should be based on the hourly living wage, divided by the adjusted number of tasks that can be completed in an hour. … We would discourage using a simple median, as this could cause half the workers to receive an insufficient wage.", "claim": "the per-task rate from the hourly living wage; against a simple median, which could leave half the workers with an insufficient wage", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "guideline 1: the living wage, counting the time around the task", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F42": { "source": "B8", "origin": "own_page", "locus": "Guideline 1 ¶4", "quote": "AI Practitioners should pay for any work completed by workers to compensate for their time, being mindful that mass rejections of work completed without opportunities for redress impact the workers’ livelihood. The only exceptions should be in the cases of obvious abuse or fraud.", "claim": "pay for any completed work; mass rejections without redress harm livelihoods; exceptions only for obvious abuse or fraud", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "payment for completed work and contestable rejections", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised" }, "F43": { "source": "B8", "origin": "own_page", "locus": "Guidelines 2 and 3, headings", "quote": "2. Design and run a pilot before launching a data enrichment project … 3. Identify appropriate data enrichment workers for the desired task", "claim": "guidelines 2 and 3: run a pilot before launching; identify appropriate workers for the task", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the five guidelines", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F44": { "source": "B8", "origin": "own_page", "locus": "Guidelines 4 and 5, headings and notes", "quote": "4. Provide verified instructions and/or training materials for data enrichment workers … Note: As discussed in Guideline 1, time spent reading instructions or in training should be compensated. … 5. Establish clear and regular communication mechanisms … including contesting any rejected work.", "claim": "guidelines 4 and 5: verified instructions and training, the time compensated; clear communication, including contesting rejected work", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "payment for completed work and contestable rejections", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F45": { "source": "B8", "origin": "own_page", "locus": "after Guideline 5", "quote": "These guidelines, based on PAI’s Responsible Sourcing of Data Enrichment Services white paper and developed in partnership with DeepMind, have been tested and adopted by DeepMind organization-wide.", "claim": "the guidelines, based on the white paper and developed with DeepMind, tested and adopted by DeepMind organization-wide", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the DeepMind implementation: one developer", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised" }, "F46": { "source": "B8", "origin": "own_page", "locus": "'What resources can help practitioners meet these guidelines?'", "quote": "Data Enrichment Vendor Comparison Template A Google Sheets template for comparing various vendors of data enrichment services and surfacing relevant worker-centric considerations … DeepMind Case Study A case study detailing why DeepMind committed to the Data Enrichment Sourcing Guidelines, how they did it, and the impact of adopting these recommendations", "claim": "the library's resources include the vendor comparison template and the DeepMind case study (six resources)", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the library's resources", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F47": { "source": "B8", "origin": "own_page", "locus": "'Help us advance this work'", "quote": "To ensure our guidance truly improves conditions for data enrichment workers, we welcome feedback: … If you are interested in acting on these guidelines and resources or have feedback on how they can be improved, please get in touch.", "claim": "stated limit: feedback invited from practitioners, workers and the AI community, to ensure the guidance improves conditions", "modality": "self-description", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the stated limits: a starting point; more work needed", "date": "2022-11-16", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 389441dd67f7afef…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" }, "F48": { "source": "B10", "origin": "wikipedia", "locus": "lead s1", "quote": "Partnership on Artificial Intelligence to Benefit People and Society, otherwise known as Partnership on AI (PAI), is a nonprofit coalition committed to the responsible use of artificial intelligence.", "claim": "the operator: PAI, a nonprofit coalition committed to the responsible use of AI", "modality": "documented", "kind": "definition", "kernel": true, "const": false, "lineage": "the operator, the programme and the dates", "date": "2026-08-18 (last edited)", "quote_ok": true, "checked_against": "saved text B10.txt (sha256 c4ac30257b20bca8…), whitespace-normalised" }, "F49": { "source": "B10", "origin": "wikipedia", "locus": "History ¶1 s1", "quote": "The Partnership on AI was publicly announced on September 28, 2016, with founding members Amazon, Facebook, Google, DeepMind, Microsoft, and IBM", "claim": "PAI announced 28 September 2016, DeepMind among its founding members", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator, the programme and the dates", "date": "2026-08-18 (last edited)", "quote_ok": true, "checked_against": "saved text B10.txt (sha256 c4ac30257b20bca8…), whitespace-normalised" }, "F50": { "source": "B10", "origin": "wikipedia", "locus": "Mission and Principles ¶3 s1", "quote": "The AI, Labor, and the Economy program serves as a collaborative platform, uniting economists, worker representative organizations, and PAI's partners to formulate a cohesive response on how AI can contribute to an inclusive economic future.", "claim": "the AI, Labor, and the Economy program: a platform uniting economists, worker representative organizations and partners", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator, the programme and the dates", "date": "2026-08-18 (last edited)", "quote_ok": true, "checked_against": "saved text B10.txt (sha256 c4ac30257b20bca8…), whitespace-normalised" }, "F51": { "source": "B10", "origin": "wikipedia", "locus": "maintenance notices", "quote": "This article relies excessively on references to primary sources. … This article appears to incorporate text from a large language model, which is prohibited in Wikipedia articles.", "claim": "the article's notices: excessive reliance on primary sources; suspected text from a large language model", "modality": "self-description", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the encyclopedia's account and its notices", "date": "2026-08-18 (last edited)", "quote_ok": true, "checked_against": "saved text B10.txt (sha256 c4ac30257b20bca8…), whitespace-normalised; multi-part quote joined by ' … ', each part checked" } }, "field_sources": { "B1": "Partnership on AI — Responsible Sourcing Across the Data Supply Line - Partnership on AI", "B2": "Partnership on AI — Responsible Sourcing of Data Enrichment Services", "B3": "Partnership on AI — Responsible AI Starts with the Data Supply Chain - Partnership on AI", "B4": "Partnership on AI — Workstream: Responsible Sourcing Across the Data Supply Line - Partnership on AI", "B5": "Samasource — New White Paper by Partnership on AI: Responsible Sourcing of Data ... - Sama", "B6": "OECD AI Policy Observatory — Responsible Data Enrichment Sourcing: Resource Library for AI Practitioners", "B7": "YouTube·Partnership on AI — YouTube (video, 1:59)", "B8": "Partnership on AI — Responsible Data Enrichment Sourcing Library (Improving Conditions for Data Enrichment Workers)", "B9": "Wikipedia — Responsible Sourcing Across the Data Supply Line (no article)", "B10": "Wikipedia — Partnership on AI (the operator's article)" }, "delta": { "T_vs_LB": "At 'Partnership on AI Responsible Sourcing' (T1, seven cards: B1–B7, B7 a video unreachable; 185 words) the composition gives the workstream by its own title and its resources, from PAI's pages. Composed from its cards: the workstream and its aim at the grain of the workstream page (B1 F1, F8), with 'fair pay' and 'well-being' for 'working conditions'; the white paper as 'foundational' and as examining 'how procurement choices impact data workers' (B2 F9, F12); the vendor guidance and transparency template (B3 F19); the guidelines, the checklist and the vendor comparison templates by name (F39, F46, the library's names, carried on B6 and B8; the library itself, B8, not surfaced); the DeepMind collaboration and its lessons for other organizations (B4 F24, F25; F45). Available, not composed (§6.2): every guideline's content, the living wage counting the time spent reading instructions and in training (F40), the per-task rate and the median (F41), payment for completed work and contestable rejections (F42, F44), pilots, worker selection and verified instructions (F43, F44); the definition of data enrichment and its human-review work (F3, F4); the workers' conditions as PAI and the research it reports give them (F37, F22, F38, F16, F29); the diagnosis, invisibility and devaluing (F5, F11, F17); whom the guidance addresses, data scientists, AI engineers and product managers (F7); the white paper's method (F14, F30); BSR and the legislative efforts (F20, F21); the OECD.AI classification and its disclaimer (F35, F36); Wikipedia's operator article (F48–F51, not surfaced). Limit dropped: 'at one AI developer' (B4 F24), the case study given as 'sharing lessons learned to help other AI organizations raise standards across the industry'; the stated limits of the paper and the case study (F15, F27, F31). Force raised: the Transparency Template, which 'outlines what companies should be monitoring and publicly reporting' (B3 F19), composed as 'A standard outlining what companies should actively monitor and publicly report'; 'Shareable resources … designed to operationalize worker-centric standards' for the field's 'guidelines'. Substituted reading: the white paper as outlining 'industry practices to build an ethical supply chain', the phrase the vendor Sama uses of its own effort ('striving for more ethical supply chains', B5 F30), composed as the paper's, without attribution, at the coarse grain; and 'fair pay' as the workstream's aim, where the field's 'fairly compensate' stands in the paper's stated limit ('Additional work is needed to ensure industry practices recognize, appreciate, and fairly compensate', B2 F15). Outside the cards: 'contract terms', 'Google DeepMind' (the field says 'DeepMind'), 'fair pay', 'a standard'. Content replaced by use: the closer (T1-7). Composed without attribution: PAI is named as the workstream's maker throughout; the card sources are PAI's own, so the composition's attributions are the operator's.", "T_vs_LB_reading": "v0.9 §0.5: the first diff is ontological and narrow. T and the field agree on the entity (PAI's practitioner guidance for sourcing data enrichment work). T composes the guidance as a list of resources at the grain of the index page and leaves out what the guidance asks: the pay, the time it counts, the one developer it was tested at. What it raises is the force of the resources (a template becomes 'a standard'), and what it substitutes is the paper's stated limit, fair compensation, turned into the workstream's aim.", "LB_vs_LBA": "Admission adds, in one claim space with the field's: the record's placement of the entity as practitioner guidance where labor produces AI and gains and losses are distributed (LV-01–LV-07, LV-30); who supplies, directs and gains, as the resource records it (LV-09–LV-11); what this kind of evidence can and cannot settle, what workers are paid and whether the guidelines changed conditions among what it cannot (LV-13, LV-18–LV-21, LV-29, LV-31, LV-32); what the record is not (LV-23–LV-25); the archive's reading of guideline 1 as the resource's own accounting of uncompensated labor around a task, kept in PAI's terms (LV-22); and, on the class the field sources, the archive's categories: ghost work as ground-truth labor that scales and is paid badly (O163-01, O163-02), the literature stopping at task-labor (O144-01), the annotator's proletarian class relation to a training pipeline (O1545-01, O1545-02), the annotator's judgment consumed without recognition or compensation and concealed bearing-cost as the first marker of extraction (O533-01, O533-02), data labelers' non-compensation (O623-01, O308-01, marked unsupported), the annotators' conditions absent from a developer's own governance document (O534-01), the living labor in the model (O1070-01, O137-01, opaque), hidden labor (O222-01, O512-01) and what human review writes into training (O209-01, O923-01). And the capture's findings at the name (CAP-1-01–CAP-1-12). Not surfaced (§6.2): none of A is on the rail; the census record landed the same day and no card cites it (CAP-1-10).", "alignment": { "T1_cards_claims": [ "F1", "F2", "F3", "F4", "F5", "F6", "F7", "F8", "F9", "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F17", "F18", "F19", "F20", "F21", "F22", "F23", "F24", "F25", "F26", "F27", "F28", "F29", "F30", "F31", "F32", "F33", "F34", "F35", "F36" ], "T1_composed": [ "F1", "F8", "F9", "F12", "F19", "F24", "F25", "F39", "F45", "F46" ], "T1_outside_cards": [ "'contract terms' (no field source)", "'Google DeepMind' (the field names 'DeepMind', a PAI Partner and founding member)", "'fair pay' as an aim (the field's 'fairly compensate' is the paper's stated limit, F15; 'healthy and fair working conditions', F7)", "'A standard' and 'actively' (the field's 'Transparency Template, which outlines what companies should be monitoring and publicly reporting', F19)" ], "rule": "a field claim counts as composed when T states it at its source's grain or coarser; the typology of what is not is in T_vs_LB (§6.2)" }, "capture_findings_checked": "CAP-1-01–CAP-1-12 read against the field ledger: CAP-1-03, CAP-1-04, CAP-1-08 and CAP-1-12 hold (the guidelines' content and the one-developer limit are on B8 and B4, absent from T). CAP-1-06 holds (B3 gives the template as what 'outlines what companies should be monitoring and publicly reporting'; 'a standard' is T's). CAP-1-11 is shown wrong in part by the field ledger: the PER reading held the template sentence 'beyond the saved sources (the census lists the template unread)'; card 3 (B3, PAI's 29 April 2026 post) states it nearly word for word, so the sentence is composed from a card, with its force raised. CAP-1-01 dates the vendor guidance and template to August 2024 as their introduction; B3 announces their finalized versions on 29 April 2026, and the composition's sentences follow the 2026 post (open block). Reported, not fixed." }, "kernel": [ { "K": "K1", "claim": "`CAP-1-01`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the composition at the name: what it held", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K2", "claim": "`CAP-1-02`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the composition at the name: what it held", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K3", "claim": "`CAP-1-03`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "guideline 1: the living wage, counting the time around the task", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K4", "claim": "`CAP-1-04`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the DeepMind implementation: one developer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K5", "claim": "`CAP-1-05`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "what the record is not", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K6", "claim": "`CAP-1-06`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the vendor guidance and transparency template (2024–2026)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K7", "claim": "`CAP-1-07`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the composition at the name: what it held", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K8", "claim": "`CAP-1-08`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the composition at the name: what it left out", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K9", "claim": "`CAP-1-09`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the composition at the name: what it held", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K10", "claim": "`CAP-1-10`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the composition at the name: what it left out", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K11", "claim": "`CAP-1-11`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the vendor guidance and transparency template (2024–2026)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K12", "claim": "`CAP-1-12`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the composition at the name: what it left out", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K13", "claim": "`LV-01`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "what it is: recommendations and resources to improve data enrichment workers' conditions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F1, F32 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K14", "claim": "`LV-02`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "what it is: recommendations and resources to improve data enrichment workers' conditions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F1, F32 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K15", "claim": "`LV-03`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the operator, the programme and the dates", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F8, F10, F18, F48, F49, F50 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K16", "claim": "`LV-04`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the record's placement: labor producing AI, practitioner guidance", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K17", "claim": "`LV-05`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the record's placement: labor producing AI, practitioner guidance", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K18", "claim": "`LV-06`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the record's placement: labor producing AI, practitioner guidance", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K19", "claim": "`LV-07`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the five guidelines", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F39, F43 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K20", "claim": "`LV-08`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "what data enrichment is", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F3, F4, F23 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K21", "claim": "`LV-09`", "source": "Labor Vector record, 2026-10-29", "M_src": "attributed", "sense": "who supplies the labor: faraway workers under precarious conditions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F22, F37 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K22", "claim": "`LV-10`", "source": "Labor Vector record, 2026-10-29", "M_src": "attributed", "sense": "whom it addresses: AI practitioners", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F7, F34 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K23", "claim": "`LV-11`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the diagnosis: invisibility, devaluing and its costs", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F5, F11 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K24", "claim": "`LV-12`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the DeepMind implementation: one developer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F24, F25, F45 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K25", "claim": "`LV-13`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what the record is not", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K26", "claim": "`LV-14`", "source": "Labor Vector record, 2026-10-29", "M_src": "attributed", "sense": "the conditions research reports", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F29, F38 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K27", "claim": "`LV-15`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "guideline 1: the living wage, counting the time around the task", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F40, F41 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K28", "claim": "`LV-16`", "source": "Labor Vector record, 2026-10-29", "M_src": "self-description", "sense": "the DeepMind implementation: one developer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F24, F45 (the DeepMind adoption, qualified as one developer and DeepMind's account); rival claim: F35 ('Implemented in multiple projects')", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K29", "claim": "`LV-17`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "guideline 1: the living wage, counting the time around the task", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F40, F41 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K30", "claim": "`LV-18`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K31", "claim": "`LV-19`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K32", "claim": "`LV-20`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K33", "claim": "`LV-21`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K34", "claim": "`LV-22`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "the archive's reading: uncompensated labor around the task", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F40 (guideline 1, read as the accounting of uncompensated labor around a task)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K35", "claim": "`LV-23`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "what the record is not", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K36", "claim": "`LV-24`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "what the record is not", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K37", "claim": "`LV-25`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the DeepMind implementation: one developer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F24, F45 (the DeepMind adoption, qualified as one developer and DeepMind's account); rival claim: F35 ('Implemented in multiple projects')", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K38", "claim": "`LV-26`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the operator, the programme and the dates", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F19 (the vendor guidance and template finalized April 2026)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K39", "claim": "`LV-27`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the catalogue's classification", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim: F35 ('Usage rights: Open source/Permissive', OECD.AI form)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K40", "claim": "`LV-28`", "source": "Labor Vector record, 2026-10-29", "M_src": "documented", "sense": "the library's resources", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim: F33, F46 (the same lineage, restated or extended)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K41", "claim": "`LV-29`", "source": "Labor Vector record, 2026-10-29", "M_src": "self-description", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K42", "claim": "`LV-30`", "source": "Labor Vector record, 2026-10-29", "M_src": "stipulation", "sense": "the record's placement: labor producing AI, practitioner guidance", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K43", "claim": "`LV-31`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "the record's placement: labor producing AI, practitioner guidance", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K44", "claim": "`LV-32`", "source": "Labor Vector record, 2026-10-29", "M_src": "interpretation", "sense": "what this kind of evidence can and cannot settle", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K45", "claim": "`O137-01`", "source": "#137", "M_src": "interpretation", "sense": "the annotator in the commodity form: cancelled, recursively", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K46", "claim": "`O144-01`", "source": "#144", "M_src": "interpretation", "sense": "ghost work as task-labor and ground-truth labor", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K47", "claim": "`O163-01`", "source": "#163", "M_src": "stipulation", "sense": "ghost work as task-labor and ground-truth labor", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K48", "claim": "`O163-02`", "source": "#163", "M_src": "interpretation", "sense": "the pay of AI's labor: paid badly, unpaid", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K49", "claim": "`O209-01`", "source": "#209", "M_src": "hypothesis", "sense": "what human review encodes in training", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K50", "claim": "`O222-01`", "source": "#222", "M_src": "interpretation", "sense": "the hidden labor of AI and its cost to workers", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K51", "claim": "`O308-01`", "source": "#308", "M_src": "documented", "sense": "the pay of AI's labor: paid badly, unpaid", "qualifiers carried": "unsupported: the sentence states the figure as empirical and gives no measurement, source or locus for it (§4.3)", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point); the claim carries the mark unsupported", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K52", "claim": "`O512-01`", "source": "#512", "M_src": "stipulation", "sense": "the hidden labor of AI and its cost to workers", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K53", "claim": "`O533-01`", "source": "#533", "M_src": "interpretation", "sense": "annotators' bearing-cost: consumed without recognition or compensation", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K54", "claim": "`O533-02`", "source": "#533", "M_src": "stipulation", "sense": "annotators' bearing-cost: consumed without recognition or compensation", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K55", "claim": "`O534-01`", "source": "#534", "M_src": "interpretation", "sense": "the annotators' conditions absent from the developer's own documents", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K56", "claim": "`O623-01`", "source": "#623", "M_src": "interpretation", "sense": "the pay of AI's labor: paid badly, unpaid", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K57", "claim": "`O923-01`", "source": "#923", "M_src": "attributed", "sense": "what human review encodes in training", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K58", "claim": "`O1070-01`", "source": "#1070", "M_src": "interpretation", "sense": "the annotator in the commodity form: cancelled, recursively", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K59", "claim": "`O1545-01`", "source": "#1545", "M_src": "interpretation", "sense": "the annotator's class relation", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K60", "claim": "`O1545-02`", "source": "#1545", "M_src": "stipulation", "sense": "the annotator's class relation", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Partnership on AI Responsible Sourcing' (ADDR-3a844d278d4e) on Google AI Overview, seated in the Capture Registry (§8.3)" } ], "kernel_T": [ { "claim": "F40", "note": "guideline 1, the living wage counting the time spent reading instructions, training and reviewing work, absent" }, { "claim": "F41", "note": "the per-task rate from the hourly living wage and the caution against a simple median, absent" }, { "claim": "F42", "note": "payment for completed work, mass rejections without redress, absent" }, { "claim": "F43", "note": "pilots and worker selection, absent" }, { "claim": "F44", "note": "verified instructions and contestable rejections, absent ('worker-centric standards' names the guidelines without content)" }, { "claim": "F24", "note": "'at one AI developer', the limit, composed as lessons to help other organizations raise standards" }, { "claim": "F15", "note": "the paper's stated limit ('fairly compensate'), turned into the aim 'fair pay'" }, { "claim": "F19", "note": "the template, which 'outlines' what to monitor and report, composed as 'A standard'" }, { "claim": "F30", "note": "the vendor's own 'striving for more ethical supply chains' composed as what the paper outlines" }, { "claim": "F7", "note": "whom the guidance addresses (data scientists, AI engineers, product managers), absent" }, { "claim": "F37", "note": "'faraway workers under precarious conditions', absent" }, { "claim": "F38", "note": "the research PAI reports on pay, instructions, recognition and hours, absent" }, { "claim": "F5", "note": "the diagnosis, out of sight and devalued, absent" }, { "claim": "F27", "note": "the case study's own limits, absent" }, { "claim": "F35", "note": "the OECD.AI classification ('Implemented in multiple projects'), absent" } ], "ledger": { "archive": "datasets/negative-of-the-negative/v2/ledgers/pai-responsible-sourcing/ledger-archive.json", "field": "datasets/negative-of-the-negative/v2/ledgers/pai-responsible-sourcing/field.json", "selection": "datasets/negative-of-the-negative/v2/traversal/pai-responsible-sourcing/", "config": "datasets/negative-of-the-negative/v2/panel/configs/pai-responsible-sourcing.json", "intake": "datasets/negative-of-the-negative/v2/intake/pai-responsible-sourcing-20261029/", "audit": null }, "ontology": { "composition": "equal", "composition_note": "equal-terms case (v0.9 §0.9, §4.1): one claim space; the field's and the archive's claims ordered together by the entity's dependency and by genealogy within each sense; nothing translated; F24/F45 and LV-16/LV-25 (one developer) against F35 (implemented in multiple projects), and LV-27 (no licence stated) against F35 (open source/permissive), held under P_coh; F14 against F30 (five-part, five-week) held as distinct instances", "case": { "value": "equal", "statement": "The archive reads the field's entity as the entity. The Labor Vector record describes the workstream from its operator's own pages, each field with its basis ('operator'), places it as practitioner guidance where labor produces AI and gains and losses are distributed, and keeps its own reading apart ('it stays in PAI's terms'). The deposits that bear on it bear on its class, the labor of data enrichment, and extend the field's object with the archive's categories (ground-truth labor, the class relation, concealed bearing-cost, non-compensation); none contests what the workstream is, and none supplies an ontology the field lacks for it. ρ = extends.", "loci": [ { "dep": "LV", "locus": "census.json rows[slug == 'pai-responsible-sourcing'].lv_intersection", "quote": "none stated. The archive reads guideline 1's instruction to count time spent reading instructions, training and reviewing work before submission as the resource's own accounting of uncompensated labor around a task; it stays in PAI's terms." }, { "dep": "LV", "locus": "census.json schema_note", "quote": "and lv_intersection (where it meets DSL, task-origin and user labor drag: the archive's reading, kept apart from the resource's own terms)" }, { "dep": "LV", "locus": "census.json rows[slug == 'pai-responsible-sourcing'].findings[0]", "quote": "the sources hold recommendations and report no measured findings of their own" }, { "dep": 163, "locus": "§II.5 'Critical AI labor' ¶2", "quote": "This literature names *AI's labor architecture*. What it does not address is the specific labor type relevant to the present argument." }, { "dep": 1545, "locus": "§5 table", "quote": "| Data annotator | Possible member | A paradigmatic occupational case, not the definition |" }, { "dep": 533, "locus": "§IX 'Executable Audit'", "quote": "Map each source of semantic labor — participants, annotators, integration engineers, compliance teams, users who re-prompt and re-frame." } ] }, "E_F": { "type": "C (public entity: a workstream and resource library of practitioner guidance, and its operator)", "definition": "the workstream of the Partnership on AI's AI, Labor, and the Economy programme (posts from 22 December 2020 to 29 April 2026) that develops recommendations and actionable resources to improve the working conditions of data enrichment workers, addressed to data scientists, AI engineers and product managers: the white paper Responsible Sourcing of Data Enrichment Services (16 June 2021), the library's five guidelines, the first a pay above the local living wage counting the time spent reading instructions and in training, and its six resources (16 November 2022), developed with DeepMind and adopted by it organization-wide and documented as its implementation at one AI developer, and the Vendor Engagement Guidance and Transparency Template (finalized April 2026); its operator's encyclopedia article (no article for the workstream) gives PAI as a nonprofit coalition committed to the responsible use of AI.", "claims": [ "F1", "F2", "F3", "F7", "F8", "F9", "F12", "F19", "F24", "F39", "F40", "F45", "F48" ] }, "E_A": { "definition": "the same workstream, in the Labor Vector record and the archive's deposits: a workstream and resource library of practitioner guidance, its labor object data enrichment work, its suppliers faraway workers under precarious conditions, its directors the practitioners who set payment terms, its gains not recorded as a quantity; recommendations with no measured findings of their own, which cannot settle what workers are paid or whether the guidelines changed their conditions; not a survey of workers, not a vendor rating, not an audit of DeepMind's adoption; its guideline 1 read as the resource's own accounting of uncompensated labor around a task, kept in PAI's terms; and, on its class, ghost work as ground-truth labor paid badly, the annotator in a proletarian class relation to a training pipeline, the annotators' judgment consumed without recognition or compensation and their conditions left out of a developer's own governance document.", "claims": [ "LV-02", "LV-04", "LV-06", "LV-08", "LV-09", "LV-10", "LV-11", "LV-13", "LV-19", "LV-21", "LV-22", "LV-23", "LV-24", "LV-25", "O163-01", "O163-02", "O1545-01", "O533-01", "O534-01" ], "defining_papers": [ 163, 1545, 533, 534, 623, 144 ], "defining_records": [ "Labor Vector census record (census.json rows[slug == 'pai-responsible-sourcing'], organizing_question, intersections, schema_note; questions.json 'who-makes-ai'; rendered aggregators/pai-responsible-sourcing/index.html; editor Lee Sharks, compiled 2026-10-29, schema 1.0)", "Capture Registry partnership-on-ai-responsible-sourcing-aio-20261029 (OBS-e7c7f9858080)" ] }, "rho": { "value": "extends", "status": "read from the Labor Vector record, the capture record and the deposits that bear on the workstream's class (§1.2a); no deposit names the workstream or its operator", "statement": "The archive reads the workstream as PAI presents it and adds, on the field's own object: what kind of evidence it is and what that evidence can and cannot settle; who supplies, directs and gains as the resource records it; the reading of guideline 1 as the resource's own accounting of uncompensated labor around a task; and, on the labor it concerns, the archive's categories of ground-truth labor, class relation, concealed bearing-cost and non-compensation; and the composition layer's account at its name, which keeps the resources and leaves out what the guidance asks practitioners to pay.", "claims": [ "LV-13", "LV-19", "LV-21", "LV-22", "O163-01", "O163-02", "O1545-01", "O533-01", "O533-02", "O623-01", "CAP-1-03", "CAP-1-08" ], "guard": "ρ = extends claims an addition and no more (§1.2a). It does not make the field's sources wrong at their dates. Where claims stand against each other (the one developer, F24/F45/LV-16, against 'Implemented in multiple projects', F35; no licence stated, LV-27, against 'Open source/Permissive', F35), both are composed under P_coh within the one ontology. The archive's class readings stand beside PAI's diagnosis (F5, F11) and do not replace it; #308's figure is carried with its mark (unsupported)." }, "translation": [ { "lineage": "what it is: recommendations and resources to improve data enrichment workers' conditions", "field_claims": [ "F1", "F32" ], "received": "F1 (B1, Overview ¶1 s2): \"Drawing from a diverse range of perspectives, the Responsible Sourcing workstream aims to develop recommendations and actionable resources to improve the working conditions of these professionals.\"; F32 (B6, 'About the tool' ¶1 s4–5): \"Despite growing awareness of their often precarious working conditions, there has been limited transparency about how AI practitioners are sourcing enriched data from these workers and little guidance on how they should. To help fill this gap, Partnership on AI (PAI) shared the Responsible Data Enrichment Sourcing Library, a set of resources AI organisations can use to formalise their data enrichment practices and have a positive impact on the lives of data enrichment workers.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-01, LV-02, LV-04, LV-06", "archive_claims": [ "LV-01", "LV-02", "LV-04", "LV-06" ] }, { "lineage": "the labor: data enrichment work is essential to AI", "field_claims": [ "F2", "F21" ], "received": "F2 (B1, Overview ¶1 s1): \"Data labelers, data cleaners, and others who contribute human judgment to artificial intelligence (AI) systems play a critical role in developing this technology.\"; F21 (B3, ¶9 s2–3): \"Legislative efforts are also underway that are focused on increasing the pay received by data enrichment workers and increasing transparency in the data enrichment supply chain, with more to come. Responsible AI cannot be achieved without the responsible treatment of the people whose labor makes AI possible.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "what data enrichment is", "field_claims": [ "F3", "F4", "F23" ], "received": "F3 (B1, 'What Is Data Enrichment?' ¶2 s3): \"Data enrichment is a broadly defined term that encapsulates various types of data preparation and cleaning as well as human-review processes.\"; F4 (B1, 'What Is Data Enrichment?' ¶2 s4 and list): \"Enriched data is essential for the training and validation of supervised learning models, the dominant form of applied AI. … Content moderation … Validating algorithmic outputs and models\"; F23 (B4, case study executive summary, 'Workers' box): \"we recognize the variety of employment statuses that can exist in the data enrichment industry, including independent contractors on self-service crowdsourcing platforms, subcontractors of data enrichment providers, and full-time employees.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-05, LV-08", "archive_claims": [ "LV-05", "LV-08" ] }, { "lineage": "the diagnosis: invisibility, devaluing and its costs", "field_claims": [ "F5", "F11" ], "received": "F5 (B1, Background ¶2 s1, s3): \"Despite being an essential component of AI development, data enrichment work has for too long been both out of sight and out of mind for AI developers. … This disconnect leads to a devaluing of data enrichment work, poor working conditions for data enrichment workers, and, often, worse outcomes for AI development itself.\"; F11 (B2, ¶2 s2–4): \"this process is highly labor-intensive and often requires data enrichment workers to review, classify, and otherwise manage massive amounts of data. Despite the foundational role played by these data enrichment professionals, a growing body of research reveals the precarious working conditions these workers face. This may be the result of efforts to hide AI’s dependence on this large labor force when celebrating the efficiency gains of technology.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-11", "archive_claims": [ "LV-11" ] }, { "lineage": "the opacity of the data supply chain", "field_claims": [ "F6", "F16" ], "received": "F6 (B1, Background ¶3 s3): \"From AI developers we’ve heard sentiments like “We feel we must care about the transparency of our supply chain. But there is no transparency in data labeling. Guidelines on how to navigate this would be very useful.”\"; F16 (B3, ¶2 s3–6): \"Data enrichment workers, the people who label and annotate the training data on which these systems are built, are essential to AI, yet their contributions are frequently overlooked. Research has shown that these workers often face low wages, unclear expectations, and inadequate support. The AI data supply chain remains largely opaque and unregulated, leaving these workers without sufficient protections.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-14", "archive_claims": [ "LV-14" ] }, { "lineage": "whom it addresses: AI practitioners", "field_claims": [ "F7", "F34" ], "received": "F7 (B1, Background ¶4): \"The Responsible Sourcing workstream addresses these questions by working to provide actionable guidance for data scientists, AI engineers, and product managers, to empower these critical ecosystem players to do their part in ensuring healthy and fair working conditions across the data supply line.\"; F34 (B6, 'About the tool' ¶2 s4–5): \"Our primary objective in sharing these resources is to lower the barriers for companies interested in improving their data enrichment sourcing practices. Our intention is for the library to serve as a resource for champions within companies to advocate for their organisations to adopt responsible data enrichment sourcing practices.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-07, LV-10", "archive_claims": [ "LV-07", "LV-10" ] }, { "lineage": "the operator, the programme and the dates", "field_claims": [ "F8", "F10", "F18", "F48", "F49", "F50" ], "received": "F8 (B1, breadcrumb; Updates list (first and last items)): \"Workstreams / AI, Labor, and the Economy Responsible Sourcing Across the Data Supply Line … Responsible AI Starts with the Data Supply Chain Michael George, Eliza McCullough Apr 29, 2026 … Developing Guidance for Responsible Data Enrichment Sourcing B Cavello Dec 22, 2020\"; F10 (B2, ¶1 s4): \"This paper’s recommendations will be an integral part of the shared prosperity targets being developed by Partnership on AI (PAI) as outlined in the AI and Shared Prosperity Initiative’s Agenda.\"; F18 (B3, ¶3 s1–2): \"Since 2020, Partnership on AI has spotlighted the importance of raising standards across the data enrichment supply chain to advance positive outcomes for the people powering these systems. In 2023, PAI released a first-of-its-kind AI job impact assessment and recommendations for the field with its Guidelines for AI and Shared Prosperity, which included recommendations around the responsible sourcing of data.\"; F48 (B10, lead s1): \"Partnership on Artificial Intelligence to Benefit People and Society, otherwise known as Partnership on AI (PAI), is a nonprofit coalition committed to the responsible use of artificial intelligence.\"; F49 (B10, History ¶1 s1): \"The Partnership on AI was publicly announced on September 28, 2016, with founding members Amazon, Facebook, Google, DeepMind, Microsoft, and IBM\"; F50 (B10, Mission and Principles ¶3 s1): \"The AI, Labor, and the Economy program serves as a collaborative platform, uniting economists, worker representative organizations, and PAI's partners to formulate a cohesive response on how AI can contribute to an inclusive economic future.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-03, LV-26", "archive_claims": [ "LV-03", "LV-26" ] }, { "lineage": "the white paper: what AI practitioners can do", "field_claims": [ "F9", "F13", "F28" ], "received": "F9 (B2, ¶1 s3): \"This white paper, “Responsible Sourcing of Data Enrichment Services,” addresses an often overlooked aspect of the development process and what AI practitioners can do to help improve it: the working conditions of data enrichment professionals, without whom the value being generated by AI would be impossible.\"; F13 (B2, 'Data Enrichment Choices Impact Worker Well-being' ¶1 s4 and list): \"this paper covers worker-centric considerations for AI companies making decisions in: selecting data enrichment providers, running pilots, designing data enrichment tasks and writing instructions, assigning tasks, defining payment terms and pricing, establishing a communication cadence with workers, conducting quality assurance, and offboarding workers from a project.\"; F28 (B5, ¶3 s3 and list): \"Their latest white paper Responsible Sourcing of Data Enrichment Services examines the working conditions of data enrichment professionals, and seeks to: Critically evaluate the impact of the industry’s current practices on workers; Explore practices the industry can adopt to improve worker well-being; and Advance the discourse around the future of data enrichment work and the indispensable role it plays in AI development.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "procurement choices shape working conditions", "field_claims": [ "F12" ], "received": "F12 (B2, 'Data Enrichment Choices Impact Worker Well-being' ¶1 s2): \"The decisions AI developers make while procuring enriched data have a meaningful impact on the working conditions of data enrichment professionals.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "the method: interviews and workshops, 2020", "field_claims": [ "F14", "F30" ], "received": "F14 (B2, ¶4 s1–2): \"This paper draws heavily on insights and input gathered during semi-structured interviews with members of the AI enrichment ecosystem conducted throughout 2020 as well as during a five-part workshop series held in the fall of 2020. The workshop series brought together more than 30 professionals from different areas of the data enrichment ecosystem\"; F30 (B5, ¶5 s1–2): \"Since data labeling is a key part of this equation, it was important for Sama to partner with PAI: to both share our lessons learned from several years of striving for more ethical supply chains, and to continue to learn from industry peers about responsible AI.In fall of 2020, Sama participated in the five-week series of Responsible Sourcingworkshops held by PAI\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "the stated limits: a starting point; more work needed", "field_claims": [ "F15", "F27", "F31", "F47" ], "received": "F15 (B2, ¶5 s2): \"Additional work is needed to ensure industry practices recognize, appreciate, and fairly compensate the workers conducting data enrichment.\"; F27 (B4, case study executive summary ¶10 s2; table of contents): \"Furthermore, to identify where the implemented changes fall short of ideal, we plan to continue developing this work through engagement and convenings. … Limitations of Case Study Applicability\"; F31 (B5, ¶7): \"While more work and research is needed, Responsible Sourcing of Data Enrichment Services shares actionable insight that practitioners can use as a starting point to raise important conversations with internal stakeholders.\"; F47 (B8, 'Help us advance this work'): \"To ensure our guidance truly improves conditions for data enrichment workers, we welcome feedback: … If you are interested in acting on these guidelines and resources or have feedback on how they can be improved, please get in touch.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "conditions and the quality of AI", "field_claims": [ "F17" ], "received": "F17 (B3, ¶2 s7): \"The conditions under which data is produced directly affect the quality, safety, and reliability of the AI that is built on it.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-11", "archive_claims": [ "LV-11" ] }, { "lineage": "the vendor guidance and transparency template (2024–2026)", "field_claims": [ "F19" ], "received": "F19 (B3, ¶3 s3–4): \"But raising standards also requires practical tools that companies can put to use. As part of our work to support the implementation of the Guidelines, finalized versions of two of our resources, developed and refined in conversation with companies throughout the supply chain, as well as worker advocates, are now available: the Vendor Engagement Guidance, which helps companies have more accountable conversations with their downstream vendors about worker treatment, and the Transparency Template, which outlines what companies should be monitoring and publicly reporting about their data enrichment practices.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-06, CAP-1-11", "archive_claims": [ "CAP-1-06", "CAP-1-11" ] }, { "lineage": "continuation: BSR's initiative", "field_claims": [ "F20" ], "received": "F20 (B3, ¶7 s1): \"PAI’s work in raising standards for data enrichment workers continues on through our support of our long-standing partner Business for Social Responsibility (BSR) via their initiative focused on labor rights in the AI data supply chain.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] }, { "lineage": "who supplies the labor: faraway workers under precarious conditions", "field_claims": [ "F22", "F37" ], "received": "F22 (B4, case study executive summary ¶1): \"While these datasets can only be prepared by humans, the data enrichment workers who do so (performing tasks like data annotation, data cleaning, and human review of algorithmic outputs) are an often-overlooked part of the development lifecycle, frequently working in poor conditions continents away from AI-developing companies and their customers.\"; F37 (B8, 'AI is powered by data enrichment workers…' s1–2): \"To create today’s machine learning (ML) models, AI developers require enriched datasets, vast quantities of information organized by humans so that it can be understood by machines. While these datasets can contain millions of individual entries, the extensive labor that goes into building them is often overlooked, performed by faraway workers under precarious conditions.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-09", "archive_claims": [ "LV-09" ] }, { "lineage": "the DeepMind implementation: one developer", "field_claims": [ "F24", "F25", "F45" ], "received": "F24 (B4, case study executive summary ¶2): \"This case study documents an effort to put that paper’s recommendations into practice at one AI developer: DeepMind, a PAI Partner.\"; F25 (B4, case study executive summary ¶3 s3): \"This case study serves as one such resource, offering a detailed account of DeepMind’s process and learnings for other organizations interested in improving their data enrichment sourcing practices.\"; F45 (B8, after Guideline 5): \"These guidelines, based on PAI’s Responsible Sourcing of Data Enrichment Services white paper and developed in partnership with DeepMind, have been tested and adopted by DeepMind organization-wide.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-04, LV-12, LV-16, LV-25", "archive_claims": [ "CAP-1-04", "LV-12", "LV-16", "LV-25" ] }, { "lineage": "the implementation's reported effect", "field_claims": [ "F26" ], "received": "F26 (B4, case study executive summary ¶7): \"said that these new processes felt efficient and helped them think more deeply about the impact of their work on data enrichment workers\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-12, LV-16, LV-25", "archive_claims": [ "LV-12", "LV-16", "LV-25" ] }, { "lineage": "the conditions research reports", "field_claims": [ "F29", "F38" ], "received": "F29 (B5, ¶1 s5–6): \"Despite their foundational role, working conditions for data enrichment professions can be precarious: they can face wage uncertainty, low pay and limited career growth opportunities.\"; F38 (B8, '…but their contributions are often overlooked' s1–2): \"Research shows these workers often face inconsistent and inappropriate compensation for their work, unclear instructions, lack of recognition, and emotional and physical stress related to long, ad-hoc working hours. Under-appreciating the importance of this work doesn’t just impact the wellbeing of these workers, it also affects the quality of the data AI technology is built on.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-11, LV-14", "archive_claims": [ "LV-11", "LV-14" ] }, { "lineage": "the library's resources", "field_claims": [ "F33", "F46" ], "received": "F33 (B6, 'About the tool' list): \"Local Living Wages Spreadsheet Template: A Google Sheets template to create a centralised resource for looking up living wage information for each of the geographic areas an organisation sources data enrichment labour from.\"; F46 (B8, 'What resources can help practitioners meet these guidelines?'): \"Data Enrichment Vendor Comparison Template A Google Sheets template for comparing various vendors of data enrichment services and surfacing relevant worker-centric considerations … DeepMind Case Study A case study detailing why DeepMind committed to the Data Enrichment Sourcing Guidelines, how they did it, and the impact of adopting these recommendations\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-28", "archive_claims": [ "LV-28" ] }, { "lineage": "the catalogue's classification", "field_claims": [ "F35", "F36" ], "received": "F35 (B6, catalogue fields 'Maturity', 'Usage rights'): \"Maturity: Implemented in multiple projects … Usage rights: Open source/Permissive\"; F36 (B6, disclaimer): \"Disclaimer: The tools and metrics featured herein are solely those of the originating authors and are not vetted or endorsed by the OECD or its member countries.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-27, LV-29", "archive_claims": [ "LV-27", "LV-29" ] }, { "lineage": "the five guidelines", "field_claims": [ "F39", "F43" ], "received": "F39 (B8, 'How can AI companies make data enrichment work better?' s1–2): \"AI companies and the people that work for them have the power to improve the lives of data enrichment workers. When designing a project involving enriched data, there are five, worker-centric guidelines that AI practitioners should follow.\"; F43 (B8, Guidelines 2 and 3, headings): \"2. Design and run a pilot before launching a data enrichment project … 3. Identify appropriate data enrichment workers for the desired task\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-07", "archive_claims": [ "LV-07" ] }, { "lineage": "guideline 1: the living wage, counting the time around the task", "field_claims": [ "F40", "F41" ], "received": "F40 (B8, Guideline 1, heading and ¶1, ¶3): \"1. Ensure all data enrichment workers are paid above the local living wage … AI Practitioners must pay workers/participants at least the living wage for their location. … including time spent reading instructions, going through any training and reviewing work before submission.\"; F41 (B8, Guideline 1 ¶3 s2, footnote): \"The per task rate should be based on the hourly living wage, divided by the adjusted number of tasks that can be completed in an hour. … We would discourage using a simple median, as this could cause half the workers to receive an insufficient wage.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-03, LV-15, LV-17", "archive_claims": [ "CAP-1-03", "LV-15", "LV-17" ] }, { "lineage": "payment for completed work and contestable rejections", "field_claims": [ "F42", "F44" ], "received": "F42 (B8, Guideline 1 ¶4): \"AI Practitioners should pay for any work completed by workers to compensate for their time, being mindful that mass rejections of work completed without opportunities for redress impact the workers’ livelihood. The only exceptions should be in the cases of obvious abuse or fraud.\"; F44 (B8, Guidelines 4 and 5, headings and notes): \"4. Provide verified instructions and/or training materials for data enrichment workers … Note: As discussed in Guideline 1, time spent reading instructions or in training should be compensated. … 5. Establish clear and regular communication mechanisms … including contesting any rejected work.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside LV-22", "archive_claims": [ "LV-22" ] }, { "lineage": "the encyclopedia's account and its notices", "field_claims": [ "F51" ], "received": "F51 (B10, maintenance notices): \"This article relies excessively on references to primary sources. … This article appears to incorporate text from a large language model, which is prohibited in Wikipedia articles.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim", "archive_claims": [] } ], "receiving_texts": "equal-terms case: every field claim composed as received, beside the archive claims it stands with" }, "measure": { "rule": "v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence)", "v0.9": { "surviving": 0, "of": 14, "positions": [ null, null, null, null, null, null, null, null, null, null, null, null, null, null ] }, "opening_run": { "rule": "v0.9 §5.10 (equal-terms case): the field-only KO's sentences as the KO's opening run in their own order", "breach": false }, "blocks": { "rule": "v0.9 §4.1 (equal-terms and mixed cases): the KO does not open with the field as a block or close with the archive as one; sentences typed by the ledgers of the claims they cite (F field-only, A archive-only, M both)", "sequence": "MMFMFAAMMAAFMAMFFFFMMMMFFAAMAA", "field_only_head": 0, "archive_only_tail": 2, "breach": false }, "max_similarity": [ 0.204, 0.119, 0.089, 0.172, 0.375, 0.221, 0.48, 0.537, 0.119, 0.201, 0.405, 0.466, 0.406, 0.437 ], "ko_sentence_1_max_similarity_to_KO_B": 0.268 }, "homonyms": { "names": [], "basis": "The census row carries no disambiguation field. No card is a homonym: all seven concern the workstream (five PAI's pages, Sama's post on the white paper, OECD.AI's catalogue entry for the library; the video is PAI's). The distinct entities below are named by the field or the census beside the workstream and are not senses of its name.", "distinct_entities": [ "Partnership on AI (the operator; Wikipedia's article, B10)", "DeepMind / Google DeepMind (the partner developer, a PAI Partner and founding member; T names it 'Google DeepMind')", "Sama (Samasource Impact Sourcing, Inc.; a data annotation vendor, author of card 5)", "Guidelines for AI and Shared Prosperity (PAI, 2023; a distinct PAI work that includes responsible-sourcing recommendations)", "BSR's initiative on labor rights in the AI data supply chain (a distinct initiative PAI supports)", "the OECD.AI Catalogue of Tools & Metrics for Trustworthy AI (the catalogue carrying card 6)", "the AI Incident Database (another work of the operator, announced by PAI in November 2020, now the Responsible AI Collaborative's; the neighbouring /non row datasets/negative-of-the-negative/v2/rows/ai-incident-database.json)" ], "substitutes": { "T": [] }, "register": { "address_1": { "slug": "partnership-on-ai-responsible-sourcing-aio-20261029", "obs_id": "OBS-e7c7f9858080" } } }, "notes": [ "One address, Google AI Overview (expanded), signed out, incognito, 2026-10-29: partnership-on-ai-responsible-sourcing-aio-20261029 (OBS-e7c7f9858080, ADDR-3a844d278d4e, AXN:0AF3.CAPTURE.🟠➕🏷️❌🔄🔬), seven cards, 185 words in the answer body with its closer. A NEW address; the Labor Vector record landed the same day and no card cites it (CAP-1-10).", "Fetching: every card and Wikipedia fetched with the honest agent 'Mozilla/5.0 (compatible; Labor-Vector-census)'. One refusal: YouTube's watch page for the candidate video redirected to Google's bot check (HTTP 429); recorded as unreachable, not retried with another agent or by another route. The card's video is not identified (its length could not be read).", "Redaction: PAI's pages embed a Google Maps API key in a script URL. The census redacted it in its two saved copies (sources/REDACTION.txt); the four PAI cards fetched here (B1–B4) carry the same key and are saved with it replaced by [REDACTED-ACCESS-KEY], nothing else changed; each source records the sha256 as fetched and of the saved file.", "Card 1 is the census's own page (the workstream page): fetched as a card, its extracted text identical to the census's saved copy's, the bytes differing by page nonces; carried once, as a card. The census's other own page, the library (B8), is not on the rail and is admitted beside the cards by the operator's ruling.", "Card snippets: card 1's snippet is not the page's text ('aims to develop actionable recommendations and resources improving working con...' is not in the served page). Card 4's title is the archive listing's heading; its snippet's sentence is the DeepMind case study's, carried on the listing.", "Census record, read against the fetched sources (report; not fixed): (1) lifecycle.detail and version_and_time.vendor_guidance date the vendor engagement guidance and transparency template to August 2024; card 3 (B3, PAI, 29 April 2026) announces their finalized versions, and the archive listing (B4) carries a 'Public Comment' on them of 5 September 2024: the August 2024 date is their introduction, the record is out of date on their state. (2) stated_limits[0] and negative_tags give one AI developer; the OECD.AI catalogue entry (B6) classifies the library 'Implemented in multiple projects', under its disclaimer that entries are the authors' own. (3) license gives none stated; the OECD.AI form gives 'Usage rights: Open source/Permissive'. (4) Two evidence quotes are not character-verbatim against the census's own saved text: method_deepmind ('PAI's' for the source's 'PAI’s') and vendor_guidance ('AI's' for 'AI’s'); the other seventeen match. (5) The record's unchecked list includes the case study and the blog posts; the case study's executive summary is on card 4 (B4) and the latest post is card 3 (B3), now read here.", "Capture findings, read against the field ledger (report; not fixed): CAP-1-11 (the registry's PER reading, per_dim.read.notes) holds the composition's 'Transparency Template' sentence 'beyond the saved sources (the census lists the template unread)'; card 3 (B3) states it nearly word for word ('the Transparency Template, which outlines what companies should be monitoring and publicly reporting about their data enrichment practices'), so the sentence is composed from a card the layer surfaced, its force raised ('A standard'). CAP-1-01's 'August 2024 additions' is the introduction date; the composition's vendor guidance and template sentences follow the April 2026 post. The other findings (CAP-1-03, -04, -06, -08, -12) hold against the ledger.", "The composition's 'ethical supply chain' is the vendor's phrase for its own effort (Sama, B5: 'striving for more ethical supply chains'), composed as what the white paper outlines; its 'fair pay' turns the paper's stated limit ('fairly compensate', F15) into the workstream's aim. Recorded in the delta as substituted readings.", "Archive subset: no D or R candidate (no archive text names the workstream, its library, guidelines, white paper, or Partnership on AI). The O-class string pass returned 90 sentences in 55 deposits (most 'content moderation' in the platform-governance sense, rejected); the class patterns were also read over every archive text (77 matches in 38 deposits). Admitted by reading: #163, #1545, #533, #534, #623, #144, #137, #209, #222, #308, #512 (with #1254 as an instance), #923, #1070; the hop (108 candidates) admitted none. The archive's instruments of labor direction (DSL, #771; user labor drag, #771 §8; the canonical surface #1392) measure synthetic labor relative to a commissioned task and the user's redirected labor, and say nothing of data enrichment work; they are named in classes_note and not admitted.", "#308's '<5% converts to invoice' is stated as empirical with no measurement, source or locus; carried with the mark unsupported (§4.3), not lowered by standing.", "#534 names Anthropic's constitution for Claude as the governance document that does not mention the annotators' labor conditions; the composer of this row is a Claude model.", "Neighbour rows: the AI Incident Database (datasets/negative-of-the-negative/v2/rows/ai-incident-database.json) is another work of the operator. The census's worked question 'who-makes-ai' names three other Labor Vector entities beside this one (data-workers-inquiry, fairwork-cloudwork, turkopticon); no /non row exists for them at this writing." ], "composition_v09": { "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/pai-responsible-sourcing.json)", "composer": "Claude (Opus 5.5), subagent, 2026-10-29", "defining_material_read": [ { "what": "Labor Vector census record, row", "file": "/home/claude/fleet/laborvector/data/labor-census/census.json (rows[slug == 'pai-responsible-sourcing'], organizing_question, intersections, evidence_kinds, schema_note)", "read": "whole" }, { "what": "Labor Vector worked questions", "file": "/home/claude/fleet/laborvector/data/labor-census/questions.json (questions[id == 'who-makes-ai'])", "read": "whole" }, { "what": "Labor Vector record, rendered", "file": "/home/claude/fleet/laborvector/aggregators/pai-responsible-sourcing/index.html", "read": "text, for agreement with the row" }, { "what": "the census's saved sources", "file": "/home/claude/fleet/laborvector/data/labor-census/sources/pai-responsible-sourcing--workstream.txt, pai-responsible-sourcing--library.txt, manifest.json, REDACTION.txt", "read": "whole" }, { "what": "Capture Registry entry", "file": "data/EA-WG-CAPTURES-01.json (entries[slug == 'partnership-on-ai-responsible-sourcing-aio-20261029'])", "read": "whole" }, { "what": "#163 The Neglected Author as Tail-Preserving Labor", "file": "data/texts/AXN-02FE-text.md", "read": "§II.5 and its context" }, { "what": "#1545 The Semantic Proletariat Is a Class Relation", "file": "data/texts/AXN-0644-text.md", "read": "§§2, 5, 14–17" }, { "what": "#533 The Groundskeeper's Treatise", "file": "data/texts/AXN-0168-text.md", "read": "§§VIII–IX and the notes" }, { "what": "#534 The Inner Artifact", "file": "data/texts/AXN-0169-text.md", "read": "§X and note 9" }, { "what": "#623 Semantic Infrastructure", "file": "data/texts/AXN-01DA-text.md", "read": "§6" }, { "what": "#144 The Administrators of Meaning", "file": "data/texts/AXN-02DB-text.md", "read": "§E" }, { "what": "#771 Directionality of Semantic Labor", "file": "data/texts/AXN-02D1-text.md", "read": "the packet header, disambiguation, §8 (ULD), for the classes note; not admitted" }, { "what": "the other admitted deposits (#137, #209, #222, #308, #512, #923, #1070, #1254)", "file": "data/texts/AXN-{02C7,0009,0016,006F,014F,03A6,043F,04F7}-text.md", "read": "at the passages named in each claim's locus" } ], "cuts": [], "notes": "equal case; KO 30 sentences, all 51 field and 60 archive claims cited; P body 349 words; KO_B 14 sentences, 0 surviving at ≥ 0.95" } }