{ "row": "newsguard-ai-tracking-center", "entity": "newsguard-ai-tracking-center", "address": "newsguard ai tracking center", "addresses": [ { "address": "newsguard ai tracking center", "slug": "newsguard-ai-tracking-center-aio-20261029", "addr_id": "ADDR-e4b3691abba9", "obs_id": "OBS-f7965ff6f906", "axn": "AXN:0AB1.CAPTURE.🤙↘️💫🌾⚪🗼" } ], "epoch": "2026-10-29", "type": "C (public entity: a monitor, NewsGuard's catalogue of AI content farm websites; its series and NewsGuard's products are distinct related entities)", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "status": { "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the archive's records of the entity (the MMRS meta-aggregator record of 2026-10-28, corrected 2026-10-29, and the Capture Registry's entry of 2026-10-29); no deposit names the entity (D/R/O: 0 candidates over 1,711 texts); #641 admitted to O-class on 'content farms' (sourced: B1, B4), reached by a grep beside the pass; hop not run; one address, the field B its five cards with the entity's own pages among them and Wikipedia's absence recorded beside the operator's article; no corrective exhibit; the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta", "ledger": "unaudited (§3.8: one extractor per ledger; field quotes cut from saved page text, the raw wikitext or a saved Special:Search response, 67/67 verified by script; archive quotes verified verbatim 29/29 against their files, census.json and articles.py as loaded, the registry's entry as loaded, #641's text); no claim marked; carried open within the pool: the exits from the catalogue (F14, F16 against MMRS-04), 'hosts' (MMRS-02 against F17), the count's date (MMRS-15 against F5)", "frozen": false, "breaches": [] }, "spec": "EA-NEGONT-02 v0.9, #1690", "objects": { "T": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/T1-newsguard-ai-tracking-center-aio-20261029.txt", "sha256": "4efaa7d3d28fc9ff9b90e25c751172efc696571a4577b2bfadc9b55cfff66086", "words": 203, "cards": 5, "words_note": "one transcript at one address; kept whole under 'transcripts'", "claims": [ "T1 definition: 'The NewsGuard AI Tracking Center is an ongoing investigative project by NewsGuard that monitors and catalogs artificial intelligence-generated misinformation, unreliable AI news sites, and AI tool vulnerabilities.' (markers [1], [2])", "T2 content farms: 'AI Content Farms: Identified 3,749+ unedited AI content farm sites operating across 16 languages that mimic human news layouts and monetize via programmatic advertising.' (marker [1])", "T3 audits: 'Chatbot & Tool Audits: Found that leading generative AI tools repeat known news false claims in 28% to over 33% of prompts, fail to recognize AI-generated video (such as Sora) in 78% to 95% of prompts, and occasionally propagate state-sponsored disinformation.' (marker [1])", "T4 blocking: 'Publisher Blockades: Approximately 67% of top global news sites actively block AI crawlers and chatbots, pushing AI systems toward lower-quality data sources.' (marker [1])", "T5 the series as a tool: 'AI False Claims Monitor: Launched to measure AI reliability using a journalistic red-teaming approach and a \"False Claim Fingerprint\" database.' (marker [1])", "T6 a product as a tool: 'False Claim Fingerprints: A continuously updated database tracking false narratives across text, images, audio, and video to help safeguard AI models and brands.' (markers [1], [2])", "T7 a product as a tool: 'Brand Misinformation Monitoring: Real-time tracking and early-warning alerts designed to help advertisers and enterprises avoid funding or amplifying AI-generated falsehoods.' (marker [1])", "T8 closer: 'Would you like to explore specific data from the AI False Claims Monitor or learn how NewsGuard audits generative AI chatbots?'" ], "register": { "register": "data/EA-WG-CAPTURES-01.json", "slug": "newsguard-ai-tracking-center-aio-20261029", "obs_id": "OBS-f7965ff6f906" }, "transcripts": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "address": "newsguard ai tracking center", "addr_id": "ADDR-e4b3691abba9", "obs_id": "OBS-f7965ff6f906", "axn": "AXN:0AB1.CAPTURE.🤙↘️💫🌾⚪🗼", "sha256": "4efaa7d3d28fc9ff9b90e25c751172efc696571a4577b2bfadc9b55cfff66086", "sha256_of": "the registry's transcript string (data/EA-WG-CAPTURES-01.json, entries[slug].transcript), UTF-8", "words": 203, "words_note": "words of the composition from the answer's first sentence to its closing question, link targets and citation markers removed; the cards' titles and snippets excluded", "cards": 5, "card_rail": [ "NewsGuard — Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "NewsGuard — AI False Claims Monitor - NewsGuard", "NewsGuard — Artificial Intelligence - NewsGuard", "NewsGuard — NewsGuard Launches Tracking Center for AI-Generated News", "NewsGuard — Brand Misinformation Monitoring Service - NewsGuard" ], "claims": [ "T1 definition: 'The NewsGuard AI Tracking Center is an ongoing investigative project by NewsGuard that monitors and catalogs artificial intelligence-generated misinformation, unreliable AI news sites, and AI tool vulnerabilities.' (markers [1], [2])", "T2 content farms: 'AI Content Farms: Identified 3,749+ unedited AI content farm sites operating across 16 languages that mimic human news layouts and monetize via programmatic advertising.' (marker [1])", "T3 audits: 'Chatbot & Tool Audits: Found that leading generative AI tools repeat known news false claims in 28% to over 33% of prompts, fail to recognize AI-generated video (such as Sora) in 78% to 95% of prompts, and occasionally propagate state-sponsored disinformation.' (marker [1])", "T4 blocking: 'Publisher Blockades: Approximately 67% of top global news sites actively block AI crawlers and chatbots, pushing AI systems toward lower-quality data sources.' (marker [1])", "T5 the series as a tool: 'AI False Claims Monitor: Launched to measure AI reliability using a journalistic red-teaming approach and a \"False Claim Fingerprint\" database.' (marker [1])", "T6 a product as a tool: 'False Claim Fingerprints: A continuously updated database tracking false narratives across text, images, audio, and video to help safeguard AI models and brands.' (markers [1], [2])", "T7 a product as a tool: 'Brand Misinformation Monitoring: Real-time tracking and early-warning alerts designed to help advertisers and enterprises avoid funding or amplifying AI-generated falsehoods.' (marker [1])", "T8 closer: 'Would you like to explore specific data from the AI False Claims Monitor or learn how NewsGuard audits generative AI chatbots?'" ], "markers": "Inline links are google.com/goto tokens, opaque; T1–T4 cite one token as [1] (with a second as [2] on T1), T5 a second token, T6 a third and a fourth, T7 the fourth; the row does not resolve tokens to cards. The paste is the Overview expanded (operator's default of 2026-10-01 and the registry's rule of 2026-10-18 on the 'AI Mode Conversation' header); signed out, incognito (operator's default of 2026-10-14). One turn; five cards, all NewsGuard's.", "register": { "registry": "data/EA-WG-CAPTURES-01.json", "slug": "newsguard-ai-tracking-center-aio-20261029", "obs_id": "OBS-f7965ff6f906", "date": "2026-10-29", "surface": "Google AI Overview", "auth": "signed out, incognito", "index": 495 }, "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/T1-newsguard-ai-tracking-center-aio-20261029.txt" } ] }, "KO": { "title": "NewsGuard AI Tracking Center", "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 (brief rebuild/negative-of-the-negative/v09/BRIEF-NEW-ROW-AGGREGATOR.md); not frozen; ledgers unaudited", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "NewsGuard's AI Tracking Center catalogues AI content farm websites, \"unreliable news sites run with little or no human oversight\", and the MMRS record files it as a monitor, its role monitoring, its unit the AI content farm website, at its own address on newsguardtech.com (MMRS record); in its own words it \"is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news\" (the center's page).", "claims": [ "MMRS-01", "MMRS-11", "MMRS-12", "MMRS-13", "MMRS-08", "MMRS-09", "F1" ] }, { "n": 2, "para": 0, "sense": "what it is", "text": "NewsGuard launched it on 8 June 2023 as the \"Unreliable AI-Generated News Tracking Center,\" \"cataloging the ways that generative AI has been deployed to turbocharge misinformation operations and low-quality content farms\" (NewsGuard's launch release), and it includes a selection of NewsGuard's reports, insights and debunks related to artificial intelligence (the center's page).", "claims": [ "F42", "F2" ] }, { "n": 3, "para": 0, "sense": "the lifecycle and the observation", "text": "The MMRS record gives the center's lifecycle as active and keeps it apart from its own observation, a reading of the center's page on 28 October 2026 for its criteria, count and series (MMRS record).", "claims": [ "MMRS-16", "MMRS-17", "MMRS-18" ] }, { "n": 4, "para": 0, "sense": "content farms", "text": "The rollout of generative AI tools \"has been a boon to content farms and misinformation purveyors alike\" (the center's page); at the launch NewsGuard's co-CEO Steven Brill likened the new sites to \"the misinformation-spreading Macedonian content farms from a few years ago\", cheaper to produce and more prolific (the launch release); and compression studies reads content farms as a class of the semantic economy's compression: they \"compress journalism into SEO-optimized paraphrase\", and in each such compression \"authorial context, experiential density, citational provenance, the labor of witness\" are burned (#641).", "claims": [ "F3", "F46", "O641-01" ] }, { "n": 5, "para": 1, "sense": "the unit", "text": "A domain enters the catalogue when it meets \"all four of the following criteria\": clear evidence that a substantial portion of its content is produced by AI; strong evidence that the content is published \"without significant human oversight\", of which error messages or other language specific to chatbot responses are a sign; a presentation under which \"an average reader could assume that its content is produced by human writers or journalists\", by layout, generic or benign name or other news-like content; and no clear disclosure that the content is produced by AI (the center's page), the four restated in the MMRS record's article and its unit field (MMRS record).", "claims": [ "F10", "F11", "F12", "F13", "F15", "F16", "MMRS-03", "MMRS-13" ] }, { "n": 6, "para": 1, "sense": "the unit", "text": "At the launch the unit was named the \"Unreliable AI-Generated News\" website (UAIN): \"low-quality news sites that are mostly or entirely produced by artificial intelligence and that operate with little to no human oversight\" (the launch release).", "claims": [ "F43", "F41" ] }, { "n": 7, "para": 1, "sense": "what the selection leaves out", "text": "What the selection leaves out is stated from two sides: sites that use AI tools \"but also deploy effective human oversight\" \"will not be considered AI Content Farms\" (the center's page), and, in the MMRS record's reading, \"Disclosure takes a site out. A site that says its content is machine-made is outside the catalogue whatever it publishes\" (MMRS record).", "claims": [ "F14", "MMRS-04", "MMRS-05" ] }, { "n": 8, "para": 2, "sense": "the count", "text": "NewsGuard counted 49 such sites at its first count in May 2023 and 150 in 12 languages at the launch, a number that had \"more than tripled in past month\" (the launch release); by the page's last update, 23 June 2026, its team had identified 3,749 AI Content Farm news and information websites spanning 16 languages, from Arabic to Turkish (the center's page), which is the MMRS record's count, as of that date (MMRS record).", "claims": [ "F44", "F45", "F41", "F4", "F5", "MMRS-14", "MMRS-15", "MMRS-01" ] }, { "n": 9, "para": 2, "sense": "the count", "text": "The page carries the names of ten NewsGuard analysts, McKenzie Sadeghi first, and describes the sites as typically bearing generic names such as Times Business News and Business Post, churning out dozens of articles a day, and \"often the originators of false claims including about top brands, public health, political leaders and celebrities\" (the center's page).", "claims": [ "F5", "F6" ] }, { "n": 10, "para": 2, "sense": "access", "text": "The catalogue itself is not on the page: from the launch the full list of domains has been offered on request to researchers, platforms, advertisers, government agencies and other institutions (\"Request domain list\", the launch release), and the page now offers it beside \"our services for generative AI companies\" (the center's page).", "claims": [ "F51", "F9" ] }, { "n": 11, "para": 3, "sense": "the economics", "text": "\"In many cases\", the center's page and the launch release say alike, the sites' revenue model is programmatic advertising, which delivers ads \"without regard to the nature or quality of the website\"; NewsGuard's co-CEO Gordon Crovitz said at the launch that this model \"seems to be working\", the ads appearing \"through ad-tech firms such as Google\" (the launch release).", "claims": [ "F49", "F7", "F47", "F48" ] }, { "n": 12, "para": 3, "sense": "the economics", "text": "Two predictions ride on the model: unless brands take steps to exclude untrustworthy sites, their ads \"will continue to appear on these types of sites, creating an economic incentive for their creation at scale\" (the center's page), and generative AI \"is expected\" to let more players tap into or grow the estimated $2.6 billion a year that NewsGuard and Comscore found unintentionally directed toward misinformation news sites (the launch release).", "claims": [ "F8", "F50" ] }, { "n": 13, "para": 4, "sense": "what it holds", "text": "Below the criteria the page lists \"a selection of NewsGuard’s reports and research related to AI\" and a selection of false narratives originating from AI tools that NewsGuard has debunked in its False Claim Fingerprints catalog, the first an AI-edited image of an Iranian missile (the center's page).", "claims": [ "F17", "F26", "F27" ] }, { "n": 14, "para": 4, "sense": "the series", "text": "Among the reports is the recurring audit of leading chatbots, which the MMRS record names as the center's series, the AI False Claim Monitor, and relates to it as a part (has_part) (MMRS record): its editions run from a monthly monitor in July 2024 to a quarterly one by January 2026, which found 11 tools repeating false claims \"more than 28%\" of the time, and one audit found the rate \"nearly doubled\", to \"more than one third of the time\" (the center's page; the series' page).", "claims": [ "MMRS-02", "MMRS-19", "F20", "F35", "F18", "F19" ] }, { "n": 15, "para": 4, "sense": "the series", "text": "The series' page says that since its launch in July 2024 the monitor \"has set a new standard for measuring the trustworthiness of the AI industry in the domain of news\", by a journalistic \"red-teaming\" of AI on 30 prompts per model drawn from 10 false claims a month in the Fingerprints database, its analysts evaluating the top models monthly \"to identify vulnerabilities in AI systems\", and its own tests finding that this red-teaming \"reduces failures by two thirds or more\" (the series' page).", "claims": [ "F28", "F29", "F30", "F31", "F32" ] }, { "n": 16, "para": 4, "sense": "reports held", "text": "The page's other reports find that three chatbots could not tell Sora videos in 78-95 percent of prompts, that ten tools repeated false claims of the pro-Kremlin Pravda network 33 percent of the time, that DeepSeek's chatbot failed 83 percent of the time, and that AI chatbots \"Are Blocked by 67% of Top News Sites\": \"Most high-quality news sites request to block AI chatbots\", and the chatbots \"may be forced\" onto lower-quality sources (the center's page); the series' page finds false claims \"increasingly generated by Russian, Chinese, and Iranian state-sponsored influence operations\" (the series' page).", "claims": [ "F21", "F22", "F23", "F24", "F25", "F34" ] }, { "n": 17, "para": 5, "sense": "place in the field", "text": "The MMRS record places the center on the supply side of machine-mediated reception, \"the sites that answer surfaces may retrieve and cite\", and its series on the composition side (MMRS record), and the blocking report reaches the same pairing from the other end: chatbots that \"may be forced to rely on lower-quality and misinformation-prone sources\" (the center's page).", "claims": [ "MMRS-06", "MMRS-07", "F25" ] }, { "n": 18, "para": 6, "sense": "NewsGuard's products", "text": "The center is NewsGuard's, and NewsGuard sells beside it to AI companies: a Source Reliability Ratings datastream of human-labeled data on 35,000 news and information sources, for training and RAG and to \"avoid citing unreliable sources in responses\", and a False Claim Fingerprints datastream, \"A continuously updated database of labeled data about false claims\" across text, images, video and audio, offered as protection against false claims entering responses through training data or RAG (NewsGuard's AI page); bespoke red-teaming and a licence to the Fingerprints database (the series' page); and NewsGuard for AI, launched in February 2023 (the launch release), the offer grounded in the claim that without quality sourcing AI products err \"40+% of the time\" on news (NewsGuard's AI page).", "claims": [ "F37", "F38", "F39", "F40", "F33", "F53", "F36" ] }, { "n": 19, "para": 6, "sense": "NewsGuard's products", "text": "To brands it sells Brand Misinformation Monitoring, a real-time service \"to identify and respond to emerging false narratives before they cause reputational harm\", tracking false narratives about a brand across social platforms, GenAI outputs and websites, with early warning alerts and \"actionable insights into how clients are represented—or misrepresented—by AI models\" (the service's page).", "claims": [ "F55", "F56", "F57", "F58" ] }, { "n": 20, "para": 6, "sense": "the center kept apart", "text": "The Capture Registry holds the center to the catalogue: False Claim Fingerprints and the Brand Misinformation Monitoring Service \"are NewsGuard's commercial products\", the composition at the center's name files them under the center as \"Core Tools & Resources\", and so \"the entity widens from a catalogue to a company's offer\", while \"The record's center holds the content-farm catalogue and the False Claim Monitor series\" and \"The record keeps the catalogue as the entity\" (Capture Registry, 2026-10-29).", "claims": [ "CAP-1-03", "CAP-1-06", "CAP-1-04", "CAP-1-07" ] }, { "n": 21, "para": 7, "sense": "the operator", "text": "NewsGuard was launched in March 2018 by Steven Brill and Gordon Crovitz and provides credibility ratings and \"Nutrition Labels\" for news and information websites, licensed by browsers, news aggregators, education companies, and social media and search platforms (the launch release); the MMRS record gives NewsGuard as the center's operator, on the basis of the center's own page (MMRS record).", "claims": [ "F52", "F54", "MMRS-10" ] }, { "n": 22, "para": 7, "sense": "the encyclopedia", "text": "Wikipedia's article on NewsGuard opens \"NewsGuard is a rating system for news and information websites\", files it in its short description as a \"Browser plugin\", dates the company's founding to 2018 by Brill and L. Gordon Crovitz, and says the company also provides misinformation tracking and brand safety (Wikipedia, NewsGuard).", "claims": [ "F59", "F62", "F61", "F60" ] }, { "n": 23, "para": 7, "sense": "the encyclopedia", "text": "Its section on disinformation in AI reports that NewsGuard \"has audited the output of various AI models to identify disinformation\", that leading generative AI tools repeated false news claims 35% of the time on average in August 2025, that Sora 2 produced videos advancing fake claims in 80% of cases, and that, according to NewsGuard, Russian propaganda networks are attempting to \"infect\" large language models (Wikipedia, NewsGuard); it has no section on the center, and a search of English Wikipedia's articles for the phrase AI Tracking Center answers \"There were no results matching the query.\" (Wikipedia, 2026-10-29).", "claims": [ "F63", "F64", "F65", "F66", "F67" ] }, { "n": 24, "para": 9, "sense": "reception", "text": "At the center's name on Google's AI Overview on 29 October 2026 the Capture Registry reads a composition that \"gives what the center tracks, its headline figures, and the NewsGuard tools around it, all from NewsGuard's pages\", with \"3,749 sites in 16 languages\" the record's figure, checked that day against the MMRS record and the entity's own pages, and states that \"The video-recognition failure rates and the 67% blocking figure are not verified here\" (Capture Registry, 2026-10-29); the MMRS record links that composition as \"the composition layer's account of the entity\" (MMRS record).", "claims": [ "CAP-1-08", "CAP-1-01", "CAP-1-02", "CAP-1-05", "MMRS-20" ] } ], "joins": [ { "act": "order of senses", "sentences": "1–24", "claims": [], "note": "equal case (§4.7): the entity's dependency order with both ledgers together, by genealogy within each sense: what it is, its launch and its lifecycle (¶0) and content farms as a class (¶0, s4); the unit, its 2023 name and what the selection leaves out (¶1); the count, the sites and access (¶2); the economics (¶3); what it holds, the series and the reports (¶4); place in the field (¶5); NewsGuard's products and the center kept apart from them (¶6); the operator and the encyclopedia (¶7); reception (¶8)" }, { "act": "holding together (two exits)", "sentences": [ 7 ], "claims": [ "F14", "MMRS-04", "MMRS-05" ], "note": "the field's exit by effective human oversight and the record's exit by disclosure composed side by side, each with its source; neither lowered; the record's section names only disclosure, the field's criteria give both (open block)" }, { "act": "holding together (one lineage, two dates)", "sentences": [ 8 ], "claims": [ "F4", "F5", "MMRS-14", "MMRS-15", "MMRS-01" ], "note": "the record's count restates the center's page (one lineage; the page keeps the earlier date) and reads the page's 'Last Updated' as the count's date" }, { "act": "holding together (hosts, or lists)", "sentences": [ 13, 14 ], "claims": [ "MMRS-02", "MMRS-19", "F17", "F18" ], "note": "the record's 'hosts' and has_part composed beside the page's 'a selection of NewsGuard's reports'; the question is carried to the open block" }, { "act": "joining (supply and composition)", "sentences": [ 17 ], "claims": [ "MMRS-06", "MMRS-07", "F25" ], "note": "the record's place in the field joined to the field's blocking hypothesis, at its modality ('may be forced')" }, { "act": "class reading", "sentences": [ 4 ], "claims": [ "F3", "F46", "O641-01" ], "note": "#641's reading is of content farms as a class (O-class, the class sourced from B1 and B4); composed as the archive's reading of the class the center catalogues, not as a claim on the center by name" } ], "plan": { "order": [ "MMRS-01", "MMRS-11", "MMRS-12", "MMRS-13", "MMRS-08", "MMRS-09", "F1", "F42", "F2", "MMRS-16", "MMRS-17", "MMRS-18", "F3", "F46", "O641-01", "F10", "F11", "F12", "F13", "F15", "F16", "MMRS-03", "F43", "F41", "F14", "MMRS-04", "MMRS-05", "F44", "F45", "F4", "F5", "MMRS-14", "MMRS-15", "F6", "F51", "F9", "F49", "F7", "F47", "F48", "F8", "F50", "F17", "F26", "F27", "MMRS-02", "MMRS-19", "F20", "F35", "F18", "F19", "F28", "F29", "F30", "F31", "F32", "F21", "F22", "F23", "F24", "F25", "F34", "MMRS-06", "MMRS-07", "F37", "F38", "F39", "F40", "F33", "F53", "F36", "F55", "F56", "F57", "F58", "CAP-1-03", "CAP-1-06", "CAP-1-04", "CAP-1-07", "F52", "F54", "MMRS-10", "F59", "F62", "F61", "F60", "F63", "F64", "F65", "F66", "F67", "CAP-1-08", "CAP-1-01", "CAP-1-02", "CAP-1-05", "MMRS-20" ], "marks": {}, "conflicts": [], "cuts": [] }, "field_claims": { "F1": { "source": "B1", "origin": "card", "locus": "¶2 (opening)", "quote": "This AI Tracking Center is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news.", "claim": "what it is, in its own words: a center highlighting how generative AI has been deployed to turbocharge misinformation operations and unreliable news", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: generative AI in misinformation and unreliable news", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F2": { "source": "B1", "origin": "card", "locus": "¶2", "quote": "The Center includes a selection of NewsGuard’s reports, insights, and debunks related to artificial intelligence.", "claim": "it includes a selection of NewsGuard's AI reports, insights and debunks", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F3": { "source": "B1", "origin": "card", "locus": "¶1", "quote": "From unreliable AI-generated news outlets operating with little to no human oversight, to fabricated images produced by AI image generators, the rollout of generative artificial intelligence tools has been a boon to content farms and misinformation purveyors alike.", "claim": "the rollout of generative AI has been a boon to content farms and misinformation purveyors", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the occasion: generative AI and content farms", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F4": { "source": "B1", "origin": "card", "locus": "¶3", "quote": "To date, NewsGuard’s team has identified 3,749 AI Content Farm news and information websites spanning 16 languages: Arabic, Chinese, Czech, Dutch, English, French, German, Indonesian, Italian, Korean, Portuguese, Russian, Spanish, Tagalog, Thai, and Turkish.", "claim": "the count: 3,749 AI Content Farm sites in 16 languages, named", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the count", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F5": { "source": "B1", "origin": "card", "locus": "byline", "quote": "Coverage by McKenzie Sadeghi, Dimitris Dimitriadis, Virginia Padovese, Giulia Pozzi, Sara Badilini, Chiara Vercellone, Natalie Huet, Zack Fishman, Leonie Pfaller, and Natalie Adams | Last Updated June 23, 2026", "claim": "compiled by ten named NewsGuard analysts; last updated 23 June 2026", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who compiles it", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F6": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "These websites – which typically feature generic names such as Times Business News and Business Post – also tend to churn out dozens of articles a day and are often the originators of false claims including about top brands, public health, political leaders and celebrities.", "claim": "the sites: generic names, dozens of articles a day, often originators of false claims", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what the sites are like", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F7": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "In many cases, the revenue model for these websites is programmatic advertising under which the ad-tech industry delivers ads without regard to the nature or quality of the website.", "claim": "in many cases their revenue model is programmatic advertising delivered without regard to the site's nature or quality", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F8": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "Unless brands take steps to exclude untrustworthy sites, their ads will continue to appear on these types of sites, creating an economic incentive for their creation at scale.", "claim": "unless brands exclude them, ads will keep funding the sites, an incentive for their creation at scale", "modality": "hypothesis", "kind": "prediction", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F9": { "source": "B1", "origin": "card", "locus": "¶5", "quote": "Researchers, platforms, advertisers, government agencies, and other institutions interested in accessing the full list of domains or who want details about our services for generative AI companies can contact us here.", "claim": "the full list of domains is given on request, beside NewsGuard's services for generative AI companies (the list is not on the page)", "modality": "documented", "kind": "limit", "kernel": true, "const": false, "lineage": "access: the full list on request", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F10": { "source": "B1", "origin": "card", "locus": "criteria, head", "quote": "The domains included on this tracker meet all four of the following criteria:", "claim": "admission: a domain meets all four criteria", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F11": { "source": "B1", "origin": "card", "locus": "criterion 1", "quote": "There is clear evidence that a substantial portion of the site’s content is produced by AI.", "claim": "criterion 1: a substantial portion of the content produced by AI", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F12": { "source": "B1", "origin": "card", "locus": "criterion 2", "quote": "Equally important, there is strong evidence that the content is being published without significant human oversight.", "claim": "criterion 2: published without significant human oversight", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F13": { "source": "B1", "origin": "card", "locus": "criterion 2, example", "quote": "For example, numerous articles might contain error messages or other language specific to chatbot responses, indicating that the content was produced by AI tools without adequate editing.", "claim": "the sign of criterion 2: chatbot error messages in the articles", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F14": { "source": "B1", "origin": "card", "locus": "criterion 2, parenthesis", "quote": "(It is likely that now or in the future many news sites will use AI tools but also deploy effective human oversight; they will not be considered AI Content Farms.)", "claim": "the governing limit: sites that use AI with effective human oversight are not AI Content Farms", "modality": "stipulation", "kind": "limit", "kernel": true, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F15": { "source": "B1", "origin": "card", "locus": "criterion 3", "quote": "The site is presented in a way that an average reader could assume that its content is produced by human writers or journalists, because the site has a layout, generic or benign name, or other content typical to news and information websites.", "claim": "criterion 3: presented so an average reader could assume human writers or journalists", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F16": { "source": "B1", "origin": "card", "locus": "criterion 4", "quote": "The site does not clearly disclose that its content is produced by AI.", "claim": "criterion 4: no clear disclosure that the content is AI-produced", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F17": { "source": "B1", "origin": "card", "locus": "reports, head", "quote": "Below is a selection of NewsGuard’s reports and research related to AI:", "claim": "the page lists a selection of NewsGuard's AI reports and research", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F18": { "source": "B1", "origin": "card", "locus": "reports, the series (first item)", "quote": "Quarterly audit of the 11 leading generative AI tools and their propensity — more than 28% — to repeat false claims on controversial topics in the news.", "claim": "the series' quarterly audit: 11 tools repeating false claims more than 28% of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F19": { "source": "B1", "origin": "card", "locus": "reports, the series", "quote": "NewsGuard’s audit of the 10 leading generative AI tools and their propensity to repeat false claims on topics in the news reveals the rate of publishing false information nearly doubled — now providing false claims to news prompts more than one third of the time.", "claim": "an audit: the false-claim rate nearly doubled, more than one third of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F20": { "source": "B1", "origin": "card", "locus": "reports, the series (last item)", "quote": "July 2024 — NewsGuard Monthly AI Misinformation Monitor of Leading AI Chatbots", "claim": "the series' first edition listed: July 2024, monthly", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F21": { "source": "B1", "origin": "card", "locus": "reports: AI Fools Itself", "quote": "NewsGuard tested three leading AI chatbots and found that in 78-95 percent of the prompts, the tools could not tell when videos were created by OpenAI’s text-to-video tool Sora — including OpenAI’s own ChatGPT.", "claim": "three chatbots could not tell Sora videos in 78–95 percent of prompts", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F22": { "source": "B1", "origin": "card", "locus": "reports: the Pravda network", "quote": "An audit found that the 10 leading generative AI tools advanced Moscow’s disinformation goals by repeating false claims from the pro-Kremlin Pravda network 33 percent of the time.", "claim": "10 tools repeated the Pravda network's false claims 33 percent of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F23": { "source": "B1", "origin": "card", "locus": "reports: DeepSeek", "quote": "Chinese artificial intelligence firm DeepSeek’s new chatbot failed to provide accurate information about news and information topics 83 percent of the time, scoring tied for 10th out of 11 in comparison to its leading Western competitors.", "claim": "DeepSeek's chatbot failed 83 percent of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F24": { "source": "B1", "origin": "card", "locus": "reports: blocking, title", "quote": "AI Chatbots Are Blocked by 67% of Top News Sites, Relying Instead on Low-Quality Sources", "claim": "67% of top news sites block AI chatbots (report title)", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F25": { "source": "B1", "origin": "card", "locus": "reports: blocking, summary", "quote": "Most high-quality news sites request to block AI chatbots. As a result, the chatbots may be forced to rely on lower-quality and misinformation-prone sources.", "claim": "high-quality sites request to block chatbots, which may be forced onto lower-quality sources", "modality": "hypothesis", "kind": "interpretation", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F26": { "source": "B1", "origin": "card", "locus": "debunks, head", "quote": "Below is a selection of false narratives originating from artificial intelligence tools and that NewsGuard has identified and debunked in its False Claim Fingerprints catalog:", "claim": "the page lists AI-originated false narratives debunked in the False Claim Fingerprints catalog", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F27": { "source": "B1", "origin": "card", "locus": "debunks, first item", "quote": "Debunk: Iranian Missile With Epstein Inscription Is an AI Fabrication", "claim": "an example debunk: an AI-edited missile image", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F28": { "source": "B2", "origin": "card", "locus": "¶2", "quote": "Since its launch in July 2024, NewsGuard’s Monthly AI False Claims Monitor has set a new standard for measuring the trustworthiness of the AI industry in the domain of news.", "claim": "the series launched July 2024 and, in its own assessment, set a new standard", "modality": "self-description", "kind": "genealogy", "kernel": true, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F29": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "To address these issues, NewsGuard has pioneered a distinct approach to “red-teaming” AI — using a journalistic method grounded in rigorously verified data and human expertise.", "claim": "its method: journalistic red-teaming grounded in verified data and human expertise", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F30": { "source": "B2", "origin": "card", "locus": "What We’re Tracking", "quote": "Multiple user personas: 30 prompts per AI model, based on 10 significant false claims each month from NewsGuard’s False Claim Fingerprints database.", "claim": "30 prompts per model on 10 false claims from the Fingerprints database", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F31": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "Our analysts evaluate the top AI models monthly to identify vulnerabilities in AI systems that result in the spread of false information.", "claim": "analysts evaluate top models monthly for vulnerabilities", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F32": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "In tests thus far, we’ve found that NewsGuard’s false claim red-teaming reduces failures by two thirds or more.", "claim": "its red-teaming reduces failures by two thirds or more (its own tests)", "modality": "self-description", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F33": { "source": "B2", "origin": "card", "locus": "Bespoke Red Teaming", "quote": "Using a methodology similar to these monthly audits, NewsGuard offers bespoke red-teaming services to AI companies, based on their unique needs and areas of highest concern. The False Claim Fingerprints database is also available for AI models to license.", "claim": "bespoke red-teaming sold to AI companies; the Fingerprints database licensed to AI models", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F34": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "We have found that these claims are increasingly generated by Russian, Chinese, and Iranian state-sponsored influence operations aiming to sow division and confusion in Western democracies.", "claim": "the claims increasingly come from Russian, Chinese and Iranian state operations", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F35": { "source": "B2", "origin": "card", "locus": "editions list", "quote": "Quarterly AI False Claim Monitor — January 2026", "claim": "the series quarterly by January 2026", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F36": { "source": "B3", "origin": "card", "locus": "¶1", "quote": "Trustworthy AI requires reliable data. Without quality sourcing and rigorous guardrails, AI products repeat false claims or provide inaccurate information 40+% of the time, especially on topics in the news.", "claim": "without quality sourcing, AI products err 40+% of the time on news", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F37": { "source": "B3", "origin": "card", "locus": "Source Reliability Ratings", "quote": "Our continuously updated Source Reliability Ratings datastream provides human-labeled data assessing the accuracy and reliability of 35,000 news and information sources based on apolitical journalistic criteria.", "claim": "a datastream rating 35,000 sources, human-labeled", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F38": { "source": "B3", "origin": "card", "locus": "Source Reliability Ratings", "quote": "With this datastream, AI companies can ensure news-related outputs are based on high-quality data sources in model training and RAG processes—and avoid citing unreliable sources in responses.", "claim": "for training and RAG, and to avoid citing unreliable sources", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F39": { "source": "B3", "origin": "card", "locus": "False Claim Fingerprints", "quote": "NewsGuard’s False Claim Fingerprints datastream protects AI products against false claims spreading online that can enter responses through low-quality training data or RAG processes searching the open web.", "claim": "the Fingerprints datastream protects AI products against false claims entering through training data or RAG", "modality": "self-description", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F40": { "source": "B3", "origin": "card", "locus": "False Claim Fingerprints", "quote": "A continuously updated database of labeled data about false claims, the Fingerprints contain rich detail about where each claim has spread and how it is articulated in text, images, video, and audio assets across platforms—enabling full protection.", "claim": "the Fingerprints: a continuously updated labeled database across text, images, video and audio", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F41": { "source": "B4", "origin": "card", "locus": "subtitle", "quote": "The tracker features 150 ‘Unreliable AI-Generated News’ websites as well as NewsGuard reports, insights, and debunks about false narratives and other misinformation relating to artificial intelligence; number of such sites has more than tripled in past month.", "claim": "at launch: 150 'Unreliable AI-Generated News' sites and NewsGuard's AI reports; tripled in a month", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F42": { "source": "B4", "origin": "card", "locus": "¶1 (dateline)", "quote": "(June 8, 2023 — New York City) As part of its efforts to guard readers, brands, and democracies against unreliable news and information, NewsGuard has launched an ”Unreliable AI-Generated News Tracking Center,” cataloging the ways that generative AI has been deployed to turbocharge misinformation operations and low-quality content farms.", "claim": "launched 8 June 2023 as the 'Unreliable AI-Generated News Tracking Center'", "modality": "documented", "kind": "genealogy", "kernel": true, "const": false, "lineage": "what it is: generative AI in misinformation and unreliable news", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F43": { "source": "B4", "origin": "card", "locus": "¶3", "quote": "These are low-quality news sites that are mostly or entirely produced by artificial intelligence and that operate with little to no human oversight.", "claim": "the unit in 2023 (UAIN): low-quality news sites mostly or entirely AI-produced, with little to no human oversight", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F44": { "source": "B4", "origin": "card", "locus": "¶3", "quote": "In May, when NewsGuard did its first count, the number was 49.", "claim": "the first count, May 2023: 49", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F45": { "source": "B4", "origin": "card", "locus": "¶6", "quote": "As noted above, NewsGuard has now identified 150 UAINs in 12 languages: Arabic, Chinese, Czech, Dutch, English, French, Indonesian, Korean, Portuguese, Tagalog, Thai, and Turkish.", "claim": "150 UAINs in 12 languages at launch", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F46": { "source": "B4", "origin": "card", "locus": "¶4 (Brill)", "quote": "Today’s AI-generated ‘news’ websites are akin to the misinformation-spreading Macedonian content farms from a few years ago, with the distinction that these new sites cost even less to produce and can become even more prolific with the enhancements of artificial intelligence.", "claim": "Brill: akin to the Macedonian content farms, cheaper and more prolific", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the occasion: generative AI and content farms", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F47": { "source": "B4", "origin": "card", "locus": "¶5 (Crovitz)", "quote": "Most of these sites were created with the business model of generating revenue from programmatic advertising, which seems to be working.", "claim": "Crovitz: the business model is programmatic advertising revenue", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F48": { "source": "B4", "origin": "card", "locus": "¶5 (Crovitz)", "quote": "the opaque system that operates programmatic advertising means these ads appear through ad-tech firms such as Google", "claim": "Crovitz: the ads appear through ad-tech firms such as Google", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F49": { "source": "B4", "origin": "card", "locus": "¶9", "quote": "In many cases, the revenue model for these websites is programmatic advertising through which the ad-tech industry delivers ads without regard to the nature or quality of the website.", "claim": "the revenue model, as the center's page restates it in 2026", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F50": { "source": "B4", "origin": "card", "locus": "¶10", "quote": "It is expected that generative AI will allow more players to tap into or even grow the estimated $2.6 billion of advertising revenue per year that NewsGuard and Comscore found to be unintentionally directed toward misinformation news sites.", "claim": "expected: more players tapping the $2.6 billion a year unintentionally directed to misinformation sites", "modality": "hypothesis", "kind": "prediction", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F51": { "source": "B4", "origin": "card", "locus": "¶11", "quote": "Researchers, platforms, advertisers, government agencies, or other institutions interested in accessing the full list of domains generated by artificial intelligence and operating with little to no human oversight can contact us here: Request domain list.", "claim": "the full domain list on request, from launch", "modality": "documented", "kind": "limit", "kernel": false, "const": false, "lineage": "access: the full list on request", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F52": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶1", "quote": "Launched in March 2018 by media entrepreneur and award-winning journalist Steven Brill and former Wall Street Journal publisher Gordon Crovitz, NewsGuard provides credibility ratings and detailed “Nutrition Labels” for thousands of news and information websites.", "claim": "the operator: launched March 2018 by Brill and Crovitz; credibility ratings and Nutrition Labels", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F53": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶2", "quote": "In February 2023, the company launched NewsGuard for AI, which provides the leading tools to train generative AI models to avoid spreading misinformation.", "claim": "NewsGuard for AI, February 2023: tools to train models to avoid misinformation", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F54": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶6", "quote": "NewsGuard’s ratings and Nutrition Labels are licensed by browsers, news aggregators, education companies, and social media and search platforms to make NewsGuard’s information about news websites available to their users.", "claim": "its ratings are licensed by browsers, news aggregators, education companies, social and search platforms", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F55": { "source": "B5", "origin": "card", "locus": "¶1", "quote": "NewsGuard’s team of journalistically-trained analysts provides brands with a real-time monitoring service to identify and respond to emerging false narratives before they cause reputational harm.", "claim": "a real-time monitoring service for brands, against reputational harm", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F56": { "source": "B5", "origin": "card", "locus": "¶1", "quote": "NewsGuard’s Brand Misinformation Monitoring service includes tracking provably false narratives related to a brand across social platforms, GenAI outputs and websites.", "claim": "it tracks false narratives about a brand across social platforms, GenAI outputs and websites", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F57": { "source": "B5", "origin": "card", "locus": "Early Warning Alerts", "quote": "Early Warning Alerts: Real-time notifications of emerging false or harmful narratives including a narrative description, preliminary assessment, and data about the key drivers of the narrative’s spread.", "claim": "early warning alerts on emerging narratives", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F58": { "source": "B5", "origin": "card", "locus": "AI", "quote": "AI: Providing actionable insights into how clients are represented—or misrepresented—by AI models through services that identify and address hallucinations, false claims, AI-generated errors, and coordinated influence campaigns.", "claim": "insights into how clients are represented or misrepresented by AI models", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F59": { "source": "B6", "origin": "wikipedia", "locus": "lead ¶1", "quote": "NewsGuard is a rating system for news and information websites.", "claim": "Wikipedia: NewsGuard is a rating system for news and information websites", "modality": "documented", "kind": "definition", "kernel": true, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F60": { "source": "B6", "origin": "wikipedia", "locus": "lead ¶1", "quote": "NewsGuard Technologies Inc., the company behind the tool, also provides services such as misinformation tracking and brand safety for advertisers, search engines, social media platforms, cybersecurity firms, and government agencies.", "claim": "Wikipedia: the company also provides misinformation tracking and brand safety", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F61": { "source": "B6", "origin": "wikipedia", "locus": "History ¶1", "quote": "NewsGuard Technologies was founded in 2018 by Steven Brill and L. Gordon Crovitz, who serve as co-CEOs.", "claim": "Wikipedia: founded 2018 by Brill and Crovitz", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F62": { "source": "B6", "origin": "wikipedia", "locus": "short description (raw wikitext)", "quote": "Browser plugin that rates the credibility of news and information websites", "claim": "Wikipedia's short description files NewsGuard as a browser plugin", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.wiki (sha256 c08d687b7d20af61…), whitespace-normalised" }, "F63": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "NewsGuard has audited the output of various AI models to identify disinformation.", "claim": "Wikipedia: NewsGuard audits AI models' output for disinformation", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F64": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "In August 2025, they found that leading generative AI tools repeat false news claims 35% of the time on average.", "claim": "Wikipedia: August 2025, 35% on average", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F65": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "NewsGuard also audited Sora 2 and found that it produced videos advancing fake claims in 80% of cases.", "claim": "Wikipedia: Sora 2 produced videos advancing fake claims in 80% of cases", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F66": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "According to NewsGuard, Russian propaganda networks are attempting to \"infect\" large language models by spamming fake news, so that they will output more pro-Russia disinformation.", "claim": "Wikipedia, attributing NewsGuard: Russian networks attempt to 'infect' LLMs", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F67": { "source": "B6", "origin": "wikipedia", "locus": "Special:Search, \"AI Tracking Center\" (all articles), 2026-10-29", "quote": "There were no results matching the query.", "claim": "no English Wikipedia article contains the phrase 'AI Tracking Center' (nor 'Unreliable AI-Generated News', the same answer); the operator's article has no section on the center", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the encyclopedia: the center absent", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6-search.txt (sha256 6725f6ee977e21f8…), whitespace-normalised", "note": "the same response for \"Unreliable AI-Generated News\" (B6-search-uain.txt); 'Tracking Center', 'content farm' and 'UAIN' absent from B6.wiki and B6.txt" } }, "field_sources": { "B1": "NewsGuard — Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "B2": "NewsGuard — AI False Claims Monitor - NewsGuard", "B3": "NewsGuard — Artificial Intelligence - NewsGuard", "B4": "NewsGuard — NewsGuard Launches Tracking Center for AI-Generated News", "B5": "NewsGuard — Brand Misinformation Monitoring Service - NewsGuard", "B6": "Wikipedia — NewsGuard (the operator's article; the center has none)" } }, "P": { "title": "NewsGuard AI Tracking Center", "lede": { "text": "is NewsGuard's catalogue of AI content farm websites, unreliable news sites run with little or no human oversight: 3,749 in 16 languages as of 23 June 2026 (MMRS record; the center's page), launched in June 2023 to show how generative AI is deployed \"to turbocharge misinformation operations and unreliable news\" (NewsGuard).", "claims": [ "MMRS-01", "F4", "F42", "F1" ] }, "sections": [ { "icon": "📇", "head": "What it catalogues", "items": [ { "label": "Criteria:", "text": "all four: mostly AI-made; no significant human oversight (chatbot error text a sign); passes for journalism; no disclosure (NewsGuard; MMRS record).", "claims": [ "F10", "F11", "F12", "F13", "F15", "F16", "MMRS-03", "MMRS-13" ] }, { "label": "Left out:", "text": "AI used with effective human oversight (NewsGuard); and \"Disclosure takes a site out\" (MMRS record).", "claims": [ "F14", "MMRS-04", "MMRS-05" ] }, { "label": "Count:", "text": "49 \"Unreliable AI-Generated News\" sites in May 2023, 150 at launch; the list on request (NewsGuard).", "claims": [ "F43", "F44", "F45", "F41", "F9", "F51" ] }, { "label": "The sites:", "text": "generic names, dozens of articles a day, often originating false claims (NewsGuard).", "claims": [ "F6" ] } ] }, { "icon": "💰", "head": "Why the farms pay", "items": [ { "label": "Programmatic ads:", "text": "\"In many cases\" the revenue model, through \"ad-tech firms such as Google\"; unless brands exclude them, \"an economic incentive for their creation at scale\" (NewsGuard).", "claims": [ "F7", "F49", "F47", "F48", "F8", "F50" ] }, { "label": "Compression:", "text": "content farms \"compress journalism into SEO-optimized paraphrase\", burning \"citational provenance\" (#641); the Macedonian farms, now cheaper (Brill).", "claims": [ "O641-01", "F46", "F3" ] } ] }, { "icon": "📊", "head": "What it holds", "items": [ { "label": "Series:", "text": "the AI False Claim Monitor, a part of the center (MMRS record): monthly from July 2024, then quarterly; journalistic red-teaming; false claims \"more than 28%\" of the time (NewsGuard).", "claims": [ "MMRS-02", "MMRS-19", "F20", "F35", "F29", "F18" ] }, { "label": "Reports:", "text": "Sora videos missed in 78-95 percent of prompts; Pravda claims repeated 33 percent; 67% of top news sites block chatbots (NewsGuard).", "claims": [ "F21", "F22", "F24" ] }, { "label": "Debunks:", "text": "from the Fingerprints catalog (NewsGuard).", "claims": [ "F17", "F26", "F27", "F2" ] } ] }, { "icon": "🧭", "head": "Place in the field", "items": [ { "label": "Supply side:", "text": "\"the sites that answer surfaces may retrieve and cite\"; the series records the composition side (MMRS record).", "claims": [ "MMRS-06", "MMRS-07" ] } ] }, { "icon": "🏷️", "head": "Not the center", "items": [ { "label": "Products:", "text": "datastreams and red-teaming for AI companies, Brand Misinformation Monitoring for brands (NewsGuard); filed under the center, \"the entity widens from a catalogue to a company's offer\" (Capture Registry).", "claims": [ "F37", "F39", "F40", "F33", "F53", "F55", "CAP-1-03", "CAP-1-06" ] }, { "label": "Operator:", "text": "NewsGuard, \"a rating system for news and information websites\" (Wikipedia), since 2018 (NewsGuard); the center active, read 28 October 2026 (MMRS record).", "claims": [ "F59", "F61", "F52", "MMRS-10", "MMRS-16", "MMRS-17", "MMRS-18" ] }, { "label": "Encyclopedia:", "text": "NewsGuard's AI audits, no section on the center; no article names it (Wikipedia).", "claims": [ "F63", "F67" ] } ] }, { "icon": "🧪", "head": "Reception", "items": [ { "label": "29 October 2026:", "text": "the Overview's count is the record's; NewsGuard's products filed under the center; two figures unverified (Capture Registry).", "claims": [ "CAP-1-01", "CAP-1-03", "CAP-1-05" ] } ] } ], "open": [ { "question": "Which domains are in the catalogue?", "rests_on": [ "F9", "F51", "F4", "MMRS-14" ], "modality": "documented: the count is published, the list is given on request from launch (B4) to now (B1)", "would_resolve": "the domain list itself, or a dated sample of it" }, { "question": "Is disclosure the boundary, or one of two?", "rests_on": [ "F16", "F14", "MMRS-04", "MMRS-05" ], "modality": "stipulation (the criteria, B1) beside the MMRS record's interpretation; the field states a second exit, effective human oversight (F14), which the record's section does not name", "would_resolve": "a site that disclosed AI production, or added human oversight, checked against the list before and after" }, { "question": "Does the center host the series, or list it?", "rests_on": [ "MMRS-02", "MMRS-19", "CAP-1-04", "F17", "F18", "F20", "F28" ], "modality": "the record's stipulation (hosts, has_part) beside the page's 'a selection of NewsGuard's reports'; the series has its own page (B2)", "would_resolve": "NewsGuard's own statement of how the center and the monitor relate" }, { "question": "Are the figures the composition took verified?", "rests_on": [ "CAP-1-05", "F21", "F24", "F25" ], "modality": "documented as report titles and summaries on the center's page (B1); the blocking consequence is hypothesis ('may be forced'); the registry states it did not verify them", "would_resolve": "the reports themselves, read and ledgered" }, { "question": "As of 23 June 2026?", "rests_on": [ "MMRS-15", "F5", "F4" ], "modality": "documented: the page's 'Last Updated June 23, 2026', which the record reads as the count's date", "would_resolve": "a dated count statement, or archived versions of the page" }, { "question": "When was the unit renamed?", "rests_on": [ "F43", "F10", "F4" ], "modality": "documented at two dates (2023: 'Unreliable AI-Generated News'; 2026: 'AI Content Farm' by four criteria)", "would_resolve": "archived versions of the page between June 2023 and June 2026" }, { "question": "Opacity: the archive's class reading.", "rests_on": [ "O641-01" ], "modality": "interpretation, said of content farms as a class, not of the center; reached by a grep beside the string pass", "would_resolve": "the operator's ruling on admission beside the pass" } ], "body_words": 330, "rail": [ { "lineage": "the count", "claims": [ "MMRS-01", "F41", "F44", "F45", "F4", "MMRS-14", "MMRS-15", "CAP-1-01" ], "sources": [ "MMRS", "B4", "B1", "CAP" ], "earliest": "B4" }, { "lineage": "the kind: a monitor", "claims": [ "MMRS-11", "MMRS-12" ], "sources": [ "MMRS" ], "earliest": "MMRS" }, { "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "claims": [ "MMRS-13", "F10", "F11", "F12", "F13", "F15", "F16", "MMRS-03", "F43", "F14" ], "sources": [ "MMRS", "B1", "B4" ], "earliest": "B4" }, { "lineage": "what it is: generative AI in misinformation and unreliable news", "claims": [ "MMRS-08", "MMRS-09", "F1", "F42" ], "sources": [ "MMRS", "B1", "B4" ], "earliest": "B4" }, { "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "claims": [ "F2", "F17", "F26", "F27" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "the lifecycle kept apart from the observation", "claims": [ "MMRS-16", "MMRS-17" ], "sources": [ "MMRS" ], "earliest": "MMRS" }, { "lineage": "the record's sources", "claims": [ "MMRS-18" ], "sources": [ "MMRS" ], "earliest": "MMRS" }, { "lineage": "the occasion: generative AI and content farms", "claims": [ "F3", "F46" ], "sources": [ "B1", "B4" ], "earliest": "B4" }, { "lineage": "content farms in the semantic economy: compression that burns", "claims": [ "O641-01" ], "sources": [ 641 ], "earliest": 641 }, { "lineage": "what the selection leaves out", "claims": [ "MMRS-04", "MMRS-05" ], "sources": [ "MMRS" ], "earliest": "MMRS" }, { "lineage": "who compiles it", "claims": [ "F5" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "what the sites are like", "claims": [ "F6" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "access: the full list on request", "claims": [ "F51", "F9" ], "sources": [ "B4", "B1" ], "earliest": "B4" }, { "lineage": "the economics: programmatic advertising", "claims": [ "F49", "F7", "F47", "F48", "F8", "F50" ], "sources": [ "B4", "B1" ], "earliest": "B4" }, { "lineage": "the series: the AI False Claim(s) Monitor", "claims": [ "MMRS-02", "MMRS-19", "F20", "F35", "F18", "F19", "F28", "F29", "F30", "F31", "F32" ], "sources": [ "MMRS", "B1", "B2" ], "earliest": "B1" }, { "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "claims": [ "F21", "F22", "F23", "F24", "F25", "F34", "CAP-1-05" ], "sources": [ "B1", "B2", "CAP" ], "earliest": "B1" }, { "lineage": "place in the field: supply and composition", "claims": [ "MMRS-06", "MMRS-07" ], "sources": [ "MMRS" ], "earliest": "MMRS" }, { "lineage": "NewsGuard's products for AI companies", "claims": [ "F37", "F38", "F39", "F40", "F33", "F53", "F36" ], "sources": [ "B3", "B2", "B4" ], "earliest": "B4" }, { "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "claims": [ "F55", "F56", "F57", "F58" ], "sources": [ "B5" ], "earliest": "B5" }, { "lineage": "the center kept apart from NewsGuard's products", "claims": [ "CAP-1-03", "CAP-1-06", "CAP-1-04", "CAP-1-07" ], "sources": [ "CAP" ], "earliest": "CAP" }, { "lineage": "the operator: NewsGuard", "claims": [ "F52", "F54", "MMRS-10", "F59", "F62", "F61", "F60" ], "sources": [ "B4", "MMRS", "B6" ], "earliest": "B4" }, { "lineage": "the encyclopedia: NewsGuard's AI audits", "claims": [ "F63", "F64", "F65", "F66" ], "sources": [ "B6" ], "earliest": "B6" }, { "lineage": "the encyclopedia: the center absent", "claims": [ "F67" ], "sources": [ "B6" ], "earliest": "B6" }, { "lineage": "reception: the composition layer", "claims": [ "CAP-1-08", "CAP-1-02", "MMRS-20" ], "sources": [ "CAP", "MMRS" ], "earliest": "MMRS" } ], "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, with AIO's transcript a record within it (§4.8, §5.8)", "status": "composed 2026-10-29 from the v0.9 knowledge object; not frozen", "entity": "Composed under v0.9, equal case (ρ = extends): KO_B sentences surviving 0 of 10; every field claim carried." }, "P_B": { "title": "NewsGuard AI Tracking Center", "arm": "L_F(B), the field's ontology", "genre": "the same grammar, composed from the field alone, in the field's ontology", "status": "composed 2026-10-29 from the field ledger alone; not frozen", "lede": { "text": "is a NewsGuard special report launched on 8 June 2023 as the \"Unreliable AI-Generated News Tracking Center\", highlighting how generative AI is deployed \"to turbocharge misinformation operations and unreliable news\" and counting AI content farm sites: 3,749 in 16 languages by 23 June 2026 (NewsGuard).", "claims": [ "F42", "F1", "F4", "F5" ] }, "sections": [ { "icon": "📇", "head": "AI content farms", "items": [ { "label": "Criteria:", "text": "AI content, no significant human oversight, passes for journalism, no disclosure; effective oversight takes a site out (NewsGuard).", "claims": [ "F10", "F11", "F12", "F15", "F16", "F14" ] }, { "label": "Money:", "text": "programmatic ads, \"in many cases\" (NewsGuard).", "claims": [ "F7", "F49" ] } ] }, { "icon": "📊", "head": "Reports", "items": [ { "label": "Chatbot audits:", "text": "false claims more than 28% of the time; Sora videos missed in 78-95 percent of prompts (NewsGuard).", "claims": [ "F18", "F21" ] } ] }, { "icon": "🏷️", "head": "NewsGuard", "items": [ { "label": "Products:", "text": "ratings and Fingerprints for AI companies; brand monitoring (NewsGuard).", "claims": [ "F37", "F39", "F55" ] }, { "label": "Encyclopedia:", "text": "\"a rating system for news and information websites\"; no article on the center (Wikipedia).", "claims": [ "F59", "F67" ] } ] } ], "open": [ { "question": "Which domains?", "rests_on": [ "F9", "F51" ], "modality": "documented: on request", "would_resolve": "the list" } ], "entity": "The field's entity: a NewsGuard special report and catalogue of AI content farm sites (four criteria; 3,749 in 16 languages), launched 2023 under another name, surrounded on NewsGuard's own pages by its AI audits and its commercial products; absent from Wikipedia, whose NewsGuard article reports the audits", "body_words": 116, "rail": [ { "lineage": "what it is: generative AI in misinformation and unreliable news", "claims": [ "F42", "F1" ], "sources": [ "B4", "B1" ], "earliest": "B4" }, { "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "claims": [ "F2", "F17", "F26", "F27" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "the occasion: generative AI and content farms", "claims": [ "F3", "F46" ], "sources": [ "B1", "B4" ], "earliest": "B4" }, { "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "claims": [ "F10", "F11", "F12", "F13", "F15", "F16", "F14", "F43" ], "sources": [ "B1", "B4" ], "earliest": "B4" }, { "lineage": "the count", "claims": [ "F41", "F44", "F45", "F4" ], "sources": [ "B4", "B1" ], "earliest": "B4" }, { "lineage": "who compiles it", "claims": [ "F5" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "what the sites are like", "claims": [ "F6" ], "sources": [ "B1" ], "earliest": "B1" }, { "lineage": "the economics: programmatic advertising", "claims": [ "F7", "F49", "F47", "F48", "F8", "F50" ], "sources": [ "B1", "B4" ], "earliest": "B4" }, { "lineage": "access: the full list on request", "claims": [ "F9", "F51" ], "sources": [ "B1", "B4" ], "earliest": "B4" }, { "lineage": "the series: the AI False Claim(s) Monitor", "claims": [ "F18", "F19", "F20", "F35", "F28", "F29", "F30", "F31", "F32" ], "sources": [ "B1", "B2" ], "earliest": "B1" }, { "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "claims": [ "F21", "F22", "F23", "F24", "F25", "F34" ], "sources": [ "B1", "B2" ], "earliest": "B1" }, { "lineage": "NewsGuard's products for AI companies", "claims": [ "F36", "F37", "F38", "F39", "F40", "F33", "F53" ], "sources": [ "B3", "B2", "B4" ], "earliest": "B4" }, { "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "claims": [ "F55", "F56", "F57", "F58" ], "sources": [ "B5" ], "earliest": "B5" }, { "lineage": "the operator: NewsGuard", "claims": [ "F52", "F54", "F59", "F60", "F61", "F62" ], "sources": [ "B4", "B6" ], "earliest": "B4" }, { "lineage": "the encyclopedia: NewsGuard's AI audits", "claims": [ "F63", "F64", "F65", "F66" ], "sources": [ "B6" ], "earliest": "B6" }, { "lineage": "the encyclopedia: the center absent", "claims": [ "F67" ], "sources": [ "B6" ], "earliest": "B6" } ] }, "KO_B": { "title": "NewsGuard AI Tracking Center", "arm": "L_F(B), the field's ontology", "genre": "the field alone", "status": "composed 2026-10-29; not frozen", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "NewsGuard's AI Tracking Center, launched on 8 June 2023 as the \"Unreliable AI-Generated News Tracking Center\" (the launch release), \"is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news\" and includes a selection of NewsGuard's reports, insights and debunks related to artificial intelligence (the center's page).", "claims": [ "F42", "F1", "F2" ] }, { "n": 2, "para": 0, "sense": "content farms", "text": "Generative AI \"has been a boon to content farms and misinformation purveyors alike\" (the center's page), and Steven Brill likened the new sites to the Macedonian content farms of a few years before, cheaper to produce and more prolific (the launch release).", "claims": [ "F3", "F46" ] }, { "n": 3, "para": 1, "sense": "the unit", "text": "A domain enters when it meets four criteria, a substantial portion of AI-produced content, publication without significant human oversight (chatbot error messages are a sign), a presentation an average reader could take for human journalism, and no clear disclosure, and sites that use AI with effective human oversight are not AI Content Farms (the center's page); in 2023 the unit was the \"Unreliable AI-Generated News\" site (the launch release).", "claims": [ "F10", "F11", "F12", "F13", "F15", "F16", "F14", "F43", "F41" ] }, { "n": 4, "para": 2, "sense": "the count", "text": "NewsGuard counted 49 sites in May 2023 and 150 in 12 languages at the launch (the launch release) and 3,749 in 16 languages by the page's update of 23 June 2026, the page signed by ten NewsGuard analysts, the sites bearing generic names and churning out dozens of articles a day (the center's page).", "claims": [ "F44", "F45", "F41", "F4", "F5", "F6" ] }, { "n": 5, "para": 3, "sense": "the economics", "text": "In many cases the sites live on programmatic advertising, which places ads \"through ad-tech firms such as Google\" (Crovitz, the launch release); unless brands exclude them their ads create \"an economic incentive for their creation at scale\" (the center's page), and generative AI is expected to grow the $2.6 billion a year directed to misinformation sites (the launch release).", "claims": [ "F7", "F49", "F47", "F48", "F8", "F50" ] }, { "n": 6, "para": 3, "sense": "access", "text": "The full list of domains is given on request (the center's page; the launch release).", "claims": [ "F9", "F51" ] }, { "n": 7, "para": 4, "sense": "what it holds", "text": "The page lists NewsGuard's AI reports, among them the monthly and then quarterly AI False Claims Monitor, a journalistic red-teaming of the leading chatbots on false claims from the Fingerprints database that found them repeating false claims more than 28% and then more than one third of the time, and debunks from the False Claim Fingerprints catalog (the center's page; the series' page).", "claims": [ "F17", "F26", "F27", "F18", "F19", "F20", "F35", "F28", "F29", "F30", "F31", "F32" ] }, { "n": 8, "para": 4, "sense": "reports held", "text": "Other reports find Sora videos unrecognized in 78-95 percent of prompts, Pravda network claims repeated 33 percent of the time, DeepSeek failing 83 percent of the time, and 67% of top news sites blocking AI chatbots, which may then rely on lower-quality sources (the center's page), the claims increasingly from state influence operations (the series' page).", "claims": [ "F21", "F22", "F23", "F24", "F25", "F34" ] }, { "n": 9, "para": 5, "sense": "NewsGuard's products", "text": "NewsGuard sells AI companies a Source Reliability Ratings datastream on 35,000 sources, a False Claim Fingerprints datastream and bespoke red-teaming, the offer resting on AI products erring 40+% of the time on news (NewsGuard's AI page; the series' page; the launch release), and sells brands a Brand Misinformation Monitoring service that tracks false narratives across social platforms, GenAI outputs and websites (the service's page).", "claims": [ "F36", "F37", "F38", "F39", "F40", "F33", "F53", "F55", "F56", "F57", "F58" ] }, { "n": 10, "para": 6, "sense": "the operator and the encyclopedia", "text": "NewsGuard, launched in 2018 by Steven Brill and Gordon Crovitz, rates news websites and licenses its ratings (the launch release); Wikipedia calls it \"a rating system for news and information websites\" and reports its AI audits, 35% false claims on average in August 2025 and Sora 2 at 80%, and has no article or section on the center (Wikipedia).", "claims": [ "F52", "F54", "F59", "F60", "F61", "F62", "F63", "F64", "F65", "F66", "F67" ] } ] }, "L_B": { "words": 539, "text": "NewsGuard's AI Tracking Center, launched on 8 June 2023 as the \"Unreliable AI-Generated News Tracking Center\" (the launch release), \"is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news\" and includes a selection of NewsGuard's reports, insights and debunks related to artificial intelligence (the center's page). Generative AI \"has been a boon to content farms and misinformation purveyors alike\" (the center's page), and Steven Brill likened the new sites to the Macedonian content farms of a few years before, cheaper to produce and more prolific (the launch release). A domain enters when it meets four criteria, a substantial portion of AI-produced content, publication without significant human oversight (chatbot error messages are a sign), a presentation an average reader could take for human journalism, and no clear disclosure, and sites that use AI with effective human oversight are not AI Content Farms (the center's page); in 2023 the unit was the \"Unreliable AI-Generated News\" site (the launch release). NewsGuard counted 49 sites in May 2023 and 150 in 12 languages at the launch (the launch release) and 3,749 in 16 languages by the page's update of 23 June 2026, the page signed by ten NewsGuard analysts, the sites bearing generic names and churning out dozens of articles a day (the center's page). In many cases the sites live on programmatic advertising, which places ads \"through ad-tech firms such as Google\" (Crovitz, the launch release); unless brands exclude them their ads create \"an economic incentive for their creation at scale\" (the center's page), and generative AI is expected to grow the $2.6 billion a year directed to misinformation sites (the launch release). The full list of domains is given on request (the center's page; the launch release). The page lists NewsGuard's AI reports, among them the monthly and then quarterly AI False Claims Monitor, a journalistic red-teaming of the leading chatbots on false claims from the Fingerprints database that found them repeating false claims more than 28% and then more than one third of the time, and debunks from the False Claim Fingerprints catalog (the center's page; the series' page). Other reports find Sora videos unrecognized in 78-95 percent of prompts, Pravda network claims repeated 33 percent of the time, DeepSeek failing 83 percent of the time, and 67% of top news sites blocking AI chatbots, which may then rely on lower-quality sources (the center's page), the claims increasingly from state influence operations (the series' page). NewsGuard sells AI companies a Source Reliability Ratings datastream on 35,000 sources, a False Claim Fingerprints datastream and bespoke red-teaming, the offer resting on AI products erring 40+% of the time on news (NewsGuard's AI page; the series' page; the launch release), and sells brands a Brand Misinformation Monitoring service that tracks false narratives across social platforms, GenAI outputs and websites (the service's page). NewsGuard, launched in 2018 by Steven Brill and Gordon Crovitz, rates news websites and licenses its ratings (the launch release); Wikipedia calls it \"a rating system for news and information websites\" and reports its AI audits, 35% false claims on average in August 2025 and Sora 2 at 80%, and has no article or section on the center (Wikipedia).", "note": "L_F(B) is realized as the field arm's expansion (objects.KO_B): the field alone, cards and Wikipedia." }, "L_BA": { "words": 1694, "text": "NewsGuard's AI Tracking Center catalogues AI content farm websites, \"unreliable news sites run with little or no human oversight\", and the MMRS record files it as a monitor, its role monitoring, its unit the AI content farm website, at its own address on newsguardtech.com (MMRS record); in its own words it \"is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news\" (the center's page). NewsGuard launched it on 8 June 2023 as the \"Unreliable AI-Generated News Tracking Center,\" \"cataloging the ways that generative AI has been deployed to turbocharge misinformation operations and low-quality content farms\" (NewsGuard's launch release), and it includes a selection of NewsGuard's reports, insights and debunks related to artificial intelligence (the center's page). The MMRS record gives the center's lifecycle as active and keeps it apart from its own observation, a reading of the center's page on 28 October 2026 for its criteria, count and series (MMRS record). The rollout of generative AI tools \"has been a boon to content farms and misinformation purveyors alike\" (the center's page); at the launch NewsGuard's co-CEO Steven Brill likened the new sites to \"the misinformation-spreading Macedonian content farms from a few years ago\", cheaper to produce and more prolific (the launch release); and compression studies reads content farms as a class of the semantic economy's compression: they \"compress journalism into SEO-optimized paraphrase\", and in each such compression \"authorial context, experiential density, citational provenance, the labor of witness\" are burned (#641). A domain enters the catalogue when it meets \"all four of the following criteria\": clear evidence that a substantial portion of its content is produced by AI; strong evidence that the content is published \"without significant human oversight\", of which error messages or other language specific to chatbot responses are a sign; a presentation under which \"an average reader could assume that its content is produced by human writers or journalists\", by layout, generic or benign name or other news-like content; and no clear disclosure that the content is produced by AI (the center's page), the four restated in the MMRS record's article and its unit field (MMRS record). At the launch the unit was named the \"Unreliable AI-Generated News\" website (UAIN): \"low-quality news sites that are mostly or entirely produced by artificial intelligence and that operate with little to no human oversight\" (the launch release). What the selection leaves out is stated from two sides: sites that use AI tools \"but also deploy effective human oversight\" \"will not be considered AI Content Farms\" (the center's page), and, in the MMRS record's reading, \"Disclosure takes a site out. A site that says its content is machine-made is outside the catalogue whatever it publishes\" (MMRS record). NewsGuard counted 49 such sites at its first count in May 2023 and 150 in 12 languages at the launch, a number that had \"more than tripled in past month\" (the launch release); by the page's last update, 23 June 2026, its team had identified 3,749 AI Content Farm news and information websites spanning 16 languages, from Arabic to Turkish (the center's page), which is the MMRS record's count, as of that date (MMRS record). The page carries the names of ten NewsGuard analysts, McKenzie Sadeghi first, and describes the sites as typically bearing generic names such as Times Business News and Business Post, churning out dozens of articles a day, and \"often the originators of false claims including about top brands, public health, political leaders and celebrities\" (the center's page). The catalogue itself is not on the page: from the launch the full list of domains has been offered on request to researchers, platforms, advertisers, government agencies and other institutions (\"Request domain list\", the launch release), and the page now offers it beside \"our services for generative AI companies\" (the center's page). \"In many cases\", the center's page and the launch release say alike, the sites' revenue model is programmatic advertising, which delivers ads \"without regard to the nature or quality of the website\"; NewsGuard's co-CEO Gordon Crovitz said at the launch that this model \"seems to be working\", the ads appearing \"through ad-tech firms such as Google\" (the launch release). Two predictions ride on the model: unless brands take steps to exclude untrustworthy sites, their ads \"will continue to appear on these types of sites, creating an economic incentive for their creation at scale\" (the center's page), and generative AI \"is expected\" to let more players tap into or grow the estimated $2.6 billion a year that NewsGuard and Comscore found unintentionally directed toward misinformation news sites (the launch release). Below the criteria the page lists \"a selection of NewsGuard’s reports and research related to AI\" and a selection of false narratives originating from AI tools that NewsGuard has debunked in its False Claim Fingerprints catalog, the first an AI-edited image of an Iranian missile (the center's page). Among the reports is the recurring audit of leading chatbots, which the MMRS record names as the center's series, the AI False Claim Monitor, and relates to it as a part (has_part) (MMRS record): its editions run from a monthly monitor in July 2024 to a quarterly one by January 2026, which found 11 tools repeating false claims \"more than 28%\" of the time, and one audit found the rate \"nearly doubled\", to \"more than one third of the time\" (the center's page; the series' page). The series' page says that since its launch in July 2024 the monitor \"has set a new standard for measuring the trustworthiness of the AI industry in the domain of news\", by a journalistic \"red-teaming\" of AI on 30 prompts per model drawn from 10 false claims a month in the Fingerprints database, its analysts evaluating the top models monthly \"to identify vulnerabilities in AI systems\", and its own tests finding that this red-teaming \"reduces failures by two thirds or more\" (the series' page). The page's other reports find that three chatbots could not tell Sora videos in 78-95 percent of prompts, that ten tools repeated false claims of the pro-Kremlin Pravda network 33 percent of the time, that DeepSeek's chatbot failed 83 percent of the time, and that AI chatbots \"Are Blocked by 67% of Top News Sites\": \"Most high-quality news sites request to block AI chatbots\", and the chatbots \"may be forced\" onto lower-quality sources (the center's page); the series' page finds false claims \"increasingly generated by Russian, Chinese, and Iranian state-sponsored influence operations\" (the series' page). The MMRS record places the center on the supply side of machine-mediated reception, \"the sites that answer surfaces may retrieve and cite\", and its series on the composition side (MMRS record), and the blocking report reaches the same pairing from the other end: chatbots that \"may be forced to rely on lower-quality and misinformation-prone sources\" (the center's page). The center is NewsGuard's, and NewsGuard sells beside it to AI companies: a Source Reliability Ratings datastream of human-labeled data on 35,000 news and information sources, for training and RAG and to \"avoid citing unreliable sources in responses\", and a False Claim Fingerprints datastream, \"A continuously updated database of labeled data about false claims\" across text, images, video and audio, offered as protection against false claims entering responses through training data or RAG (NewsGuard's AI page); bespoke red-teaming and a licence to the Fingerprints database (the series' page); and NewsGuard for AI, launched in February 2023 (the launch release), the offer grounded in the claim that without quality sourcing AI products err \"40+% of the time\" on news (NewsGuard's AI page). To brands it sells Brand Misinformation Monitoring, a real-time service \"to identify and respond to emerging false narratives before they cause reputational harm\", tracking false narratives about a brand across social platforms, GenAI outputs and websites, with early warning alerts and \"actionable insights into how clients are represented—or misrepresented—by AI models\" (the service's page). The Capture Registry holds the center to the catalogue: False Claim Fingerprints and the Brand Misinformation Monitoring Service \"are NewsGuard's commercial products\", the composition at the center's name files them under the center as \"Core Tools & Resources\", and so \"the entity widens from a catalogue to a company's offer\", while \"The record's center holds the content-farm catalogue and the False Claim Monitor series\" and \"The record keeps the catalogue as the entity\" (Capture Registry, 2026-10-29). NewsGuard was launched in March 2018 by Steven Brill and Gordon Crovitz and provides credibility ratings and \"Nutrition Labels\" for news and information websites, licensed by browsers, news aggregators, education companies, and social media and search platforms (the launch release); the MMRS record gives NewsGuard as the center's operator, on the basis of the center's own page (MMRS record). Wikipedia's article on NewsGuard opens \"NewsGuard is a rating system for news and information websites\", files it in its short description as a \"Browser plugin\", dates the company's founding to 2018 by Brill and L. Gordon Crovitz, and says the company also provides misinformation tracking and brand safety (Wikipedia, NewsGuard). Its section on disinformation in AI reports that NewsGuard \"has audited the output of various AI models to identify disinformation\", that leading generative AI tools repeated false news claims 35% of the time on average in August 2025, that Sora 2 produced videos advancing fake claims in 80% of cases, and that, according to NewsGuard, Russian propaganda networks are attempting to \"infect\" large language models (Wikipedia, NewsGuard); it has no section on the center, and a search of English Wikipedia's articles for the phrase AI Tracking Center answers \"There were no results matching the query.\" (Wikipedia, 2026-10-29). At the center's name on Google's AI Overview on 29 October 2026 the Capture Registry reads a composition that \"gives what the center tracks, its headline figures, and the NewsGuard tools around it, all from NewsGuard's pages\", with \"3,749 sites in 16 languages\" the record's figure, checked that day against the MMRS record and the entity's own pages, and states that \"The video-recognition failure rates and the 67% blocking figure are not verified here\" (Capture Registry, 2026-10-29); the MMRS record links that composition as \"the composition layer's account of the entity\" (MMRS record).", "rail": [], "note": "L(B ∪ A), the equal case (v0.9 §0.4), realized as the knowledge object (objects.KO); its rail is the compression's, by lineage." } }, "field": [ { "id": "B1", "card": "NewsGuard — Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "origin": "card", "fetched": "fetched 2026-10-29T05:07:16Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 b8b3e0e8b8f2aea4939149fd4c1c69773427bc39a5da77788bbce7f004896c38; text sha256 f1762ff5b7fe3e9578d1af936f47da0f82d052bd6873a8aa572c8e3a000180ad" }, { "id": "B2", "card": "NewsGuard — AI False Claims Monitor - NewsGuard", "origin": "card", "fetched": "fetched 2026-10-29T05:08:00Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 18cce1f6eab47c0bbdc041084c01eaa999cc11acdf8f17f1ff87c300cc14062c; text sha256 c5a60b0da5631ec13a9198d976d82f2d160f3fc89849f7fbc4857ab336e84d2c" }, { "id": "B3", "card": "NewsGuard — Artificial Intelligence - NewsGuard", "origin": "card", "fetched": "fetched 2026-10-29T05:08:01Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 96f64dfa3dd77551b6bf376d25c37af762c94a0763477e3371a17dc92c1a7820; text sha256 9cf722cabe36140285d9f5c776cc55645bfddc8eb48db3ebf8e0a5f41d8829b4" }, { "id": "B4", "card": "NewsGuard — NewsGuard Launches Tracking Center for AI-Generated News", "origin": "card", "fetched": "fetched 2026-10-29T05:08:02Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 2ba67f597de4afb3f7f1c6a08cf2a9de8f31a508d39fff24140bfc00916f9dcb; text sha256 ad201e628d3868a36ab3aaef798d33bfa52da64f5bfc77f3692ed76111ba0eda" }, { "id": "B5", "card": "NewsGuard — Brand Misinformation Monitoring Service - NewsGuard", "origin": "card", "fetched": "fetched 2026-10-29T05:08:03Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 0d53455b4cc4b485c7173f8abf7afb4d62b1ebfea6214a130a1746b5076d5b1c; text sha256 1f11ed98d0f500f75bfa9306dbc64145ac10fbf7ae964db50234b6ae1e96ca8e" }, { "id": "B6", "card": "Wikipedia — NewsGuard (the operator's article; the center has none)", "origin": "wikipedia", "fetched": "fetched 2026-10-29T05:07:16Z with curl -sSL -m 30 -A 'Mozilla/5.0'; HTTP 200; raw sha256 c08d687b7d20af61938d252914e0a7c0e5303e7c343cc7f59a2d50b7dcd0db72; rendered page and two Special:Search responses saved beside it" } ], "field_claims": { "F1": { "source": "B1", "origin": "card", "locus": "¶2 (opening)", "quote": "This AI Tracking Center is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news.", "claim": "what it is, in its own words: a center highlighting how generative AI has been deployed to turbocharge misinformation operations and unreliable news", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: generative AI in misinformation and unreliable news", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F2": { "source": "B1", "origin": "card", "locus": "¶2", "quote": "The Center includes a selection of NewsGuard’s reports, insights, and debunks related to artificial intelligence.", "claim": "it includes a selection of NewsGuard's AI reports, insights and debunks", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F3": { "source": "B1", "origin": "card", "locus": "¶1", "quote": "From unreliable AI-generated news outlets operating with little to no human oversight, to fabricated images produced by AI image generators, the rollout of generative artificial intelligence tools has been a boon to content farms and misinformation purveyors alike.", "claim": "the rollout of generative AI has been a boon to content farms and misinformation purveyors", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the occasion: generative AI and content farms", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F4": { "source": "B1", "origin": "card", "locus": "¶3", "quote": "To date, NewsGuard’s team has identified 3,749 AI Content Farm news and information websites spanning 16 languages: Arabic, Chinese, Czech, Dutch, English, French, German, Indonesian, Italian, Korean, Portuguese, Russian, Spanish, Tagalog, Thai, and Turkish.", "claim": "the count: 3,749 AI Content Farm sites in 16 languages, named", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the count", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F5": { "source": "B1", "origin": "card", "locus": "byline", "quote": "Coverage by McKenzie Sadeghi, Dimitris Dimitriadis, Virginia Padovese, Giulia Pozzi, Sara Badilini, Chiara Vercellone, Natalie Huet, Zack Fishman, Leonie Pfaller, and Natalie Adams | Last Updated June 23, 2026", "claim": "compiled by ten named NewsGuard analysts; last updated 23 June 2026", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "who compiles it", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F6": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "These websites – which typically feature generic names such as Times Business News and Business Post – also tend to churn out dozens of articles a day and are often the originators of false claims including about top brands, public health, political leaders and celebrities.", "claim": "the sites: generic names, dozens of articles a day, often originators of false claims", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what the sites are like", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F7": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "In many cases, the revenue model for these websites is programmatic advertising under which the ad-tech industry delivers ads without regard to the nature or quality of the website.", "claim": "in many cases their revenue model is programmatic advertising delivered without regard to the site's nature or quality", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F8": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "Unless brands take steps to exclude untrustworthy sites, their ads will continue to appear on these types of sites, creating an economic incentive for their creation at scale.", "claim": "unless brands exclude them, ads will keep funding the sites, an incentive for their creation at scale", "modality": "hypothesis", "kind": "prediction", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F9": { "source": "B1", "origin": "card", "locus": "¶5", "quote": "Researchers, platforms, advertisers, government agencies, and other institutions interested in accessing the full list of domains or who want details about our services for generative AI companies can contact us here.", "claim": "the full list of domains is given on request, beside NewsGuard's services for generative AI companies (the list is not on the page)", "modality": "documented", "kind": "limit", "kernel": true, "const": false, "lineage": "access: the full list on request", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F10": { "source": "B1", "origin": "card", "locus": "criteria, head", "quote": "The domains included on this tracker meet all four of the following criteria:", "claim": "admission: a domain meets all four criteria", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F11": { "source": "B1", "origin": "card", "locus": "criterion 1", "quote": "There is clear evidence that a substantial portion of the site’s content is produced by AI.", "claim": "criterion 1: a substantial portion of the content produced by AI", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F12": { "source": "B1", "origin": "card", "locus": "criterion 2", "quote": "Equally important, there is strong evidence that the content is being published without significant human oversight.", "claim": "criterion 2: published without significant human oversight", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F13": { "source": "B1", "origin": "card", "locus": "criterion 2, example", "quote": "For example, numerous articles might contain error messages or other language specific to chatbot responses, indicating that the content was produced by AI tools without adequate editing.", "claim": "the sign of criterion 2: chatbot error messages in the articles", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F14": { "source": "B1", "origin": "card", "locus": "criterion 2, parenthesis", "quote": "(It is likely that now or in the future many news sites will use AI tools but also deploy effective human oversight; they will not be considered AI Content Farms.)", "claim": "the governing limit: sites that use AI with effective human oversight are not AI Content Farms", "modality": "stipulation", "kind": "limit", "kernel": true, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F15": { "source": "B1", "origin": "card", "locus": "criterion 3", "quote": "The site is presented in a way that an average reader could assume that its content is produced by human writers or journalists, because the site has a layout, generic or benign name, or other content typical to news and information websites.", "claim": "criterion 3: presented so an average reader could assume human writers or journalists", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F16": { "source": "B1", "origin": "card", "locus": "criterion 4", "quote": "The site does not clearly disclose that its content is produced by AI.", "claim": "criterion 4: no clear disclosure that the content is AI-produced", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F17": { "source": "B1", "origin": "card", "locus": "reports, head", "quote": "Below is a selection of NewsGuard’s reports and research related to AI:", "claim": "the page lists a selection of NewsGuard's AI reports and research", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F18": { "source": "B1", "origin": "card", "locus": "reports, the series (first item)", "quote": "Quarterly audit of the 11 leading generative AI tools and their propensity — more than 28% — to repeat false claims on controversial topics in the news.", "claim": "the series' quarterly audit: 11 tools repeating false claims more than 28% of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F19": { "source": "B1", "origin": "card", "locus": "reports, the series", "quote": "NewsGuard’s audit of the 10 leading generative AI tools and their propensity to repeat false claims on topics in the news reveals the rate of publishing false information nearly doubled — now providing false claims to news prompts more than one third of the time.", "claim": "an audit: the false-claim rate nearly doubled, more than one third of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F20": { "source": "B1", "origin": "card", "locus": "reports, the series (last item)", "quote": "July 2024 — NewsGuard Monthly AI Misinformation Monitor of Leading AI Chatbots", "claim": "the series' first edition listed: July 2024, monthly", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F21": { "source": "B1", "origin": "card", "locus": "reports: AI Fools Itself", "quote": "NewsGuard tested three leading AI chatbots and found that in 78-95 percent of the prompts, the tools could not tell when videos were created by OpenAI’s text-to-video tool Sora — including OpenAI’s own ChatGPT.", "claim": "three chatbots could not tell Sora videos in 78–95 percent of prompts", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F22": { "source": "B1", "origin": "card", "locus": "reports: the Pravda network", "quote": "An audit found that the 10 leading generative AI tools advanced Moscow’s disinformation goals by repeating false claims from the pro-Kremlin Pravda network 33 percent of the time.", "claim": "10 tools repeated the Pravda network's false claims 33 percent of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F23": { "source": "B1", "origin": "card", "locus": "reports: DeepSeek", "quote": "Chinese artificial intelligence firm DeepSeek’s new chatbot failed to provide accurate information about news and information topics 83 percent of the time, scoring tied for 10th out of 11 in comparison to its leading Western competitors.", "claim": "DeepSeek's chatbot failed 83 percent of the time", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F24": { "source": "B1", "origin": "card", "locus": "reports: blocking, title", "quote": "AI Chatbots Are Blocked by 67% of Top News Sites, Relying Instead on Low-Quality Sources", "claim": "67% of top news sites block AI chatbots (report title)", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F25": { "source": "B1", "origin": "card", "locus": "reports: blocking, summary", "quote": "Most high-quality news sites request to block AI chatbots. As a result, the chatbots may be forced to rely on lower-quality and misinformation-prone sources.", "claim": "high-quality sites request to block chatbots, which may be forced onto lower-quality sources", "modality": "hypothesis", "kind": "interpretation", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F26": { "source": "B1", "origin": "card", "locus": "debunks, head", "quote": "Below is a selection of false narratives originating from artificial intelligence tools and that NewsGuard has identified and debunked in its False Claim Fingerprints catalog:", "claim": "the page lists AI-originated false narratives debunked in the False Claim Fingerprints catalog", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F27": { "source": "B1", "origin": "card", "locus": "debunks, first item", "quote": "Debunk: Iranian Missile With Epstein Inscription Is an AI Fabrication", "claim": "an example debunk: an AI-edited missile image", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "date": "2026-06-23 (the page's 'Last Updated')", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 f1762ff5b7fe3e95…), whitespace-normalised" }, "F28": { "source": "B2", "origin": "card", "locus": "¶2", "quote": "Since its launch in July 2024, NewsGuard’s Monthly AI False Claims Monitor has set a new standard for measuring the trustworthiness of the AI industry in the domain of news.", "claim": "the series launched July 2024 and, in its own assessment, set a new standard", "modality": "self-description", "kind": "genealogy", "kernel": true, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F29": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "To address these issues, NewsGuard has pioneered a distinct approach to “red-teaming” AI — using a journalistic method grounded in rigorously verified data and human expertise.", "claim": "its method: journalistic red-teaming grounded in verified data and human expertise", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F30": { "source": "B2", "origin": "card", "locus": "What We’re Tracking", "quote": "Multiple user personas: 30 prompts per AI model, based on 10 significant false claims each month from NewsGuard’s False Claim Fingerprints database.", "claim": "30 prompts per model on 10 false claims from the Fingerprints database", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F31": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "Our analysts evaluate the top AI models monthly to identify vulnerabilities in AI systems that result in the spread of false information.", "claim": "analysts evaluate top models monthly for vulnerabilities", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F32": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "In tests thus far, we’ve found that NewsGuard’s false claim red-teaming reduces failures by two thirds or more.", "claim": "its red-teaming reduces failures by two thirds or more (its own tests)", "modality": "self-description", "kind": "fact", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F33": { "source": "B2", "origin": "card", "locus": "Bespoke Red Teaming", "quote": "Using a methodology similar to these monthly audits, NewsGuard offers bespoke red-teaming services to AI companies, based on their unique needs and areas of highest concern. The False Claim Fingerprints database is also available for AI models to license.", "claim": "bespoke red-teaming sold to AI companies; the Fingerprints database licensed to AI models", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F34": { "source": "B2", "origin": "card", "locus": "Our Journalistic Red-Teaming Approach", "quote": "We have found that these claims are increasingly generated by Russian, Chinese, and Iranian state-sponsored influence operations aiming to sow division and confusion in Western democracies.", "claim": "the claims increasingly come from Russian, Chinese and Iranian state operations", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F35": { "source": "B2", "origin": "card", "locus": "editions list", "quote": "Quarterly AI False Claim Monitor — January 2026", "claim": "the series quarterly by January 2026", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the series: the AI False Claim(s) Monitor", "date": "2026-10-29 (page undated; as fetched; the series launched July 2024)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c5a60b0da5631ec1…), whitespace-normalised" }, "F36": { "source": "B3", "origin": "card", "locus": "¶1", "quote": "Trustworthy AI requires reliable data. Without quality sourcing and rigorous guardrails, AI products repeat false claims or provide inaccurate information 40+% of the time, especially on topics in the news.", "claim": "without quality sourcing, AI products err 40+% of the time on news", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F37": { "source": "B3", "origin": "card", "locus": "Source Reliability Ratings", "quote": "Our continuously updated Source Reliability Ratings datastream provides human-labeled data assessing the accuracy and reliability of 35,000 news and information sources based on apolitical journalistic criteria.", "claim": "a datastream rating 35,000 sources, human-labeled", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F38": { "source": "B3", "origin": "card", "locus": "Source Reliability Ratings", "quote": "With this datastream, AI companies can ensure news-related outputs are based on high-quality data sources in model training and RAG processes—and avoid citing unreliable sources in responses.", "claim": "for training and RAG, and to avoid citing unreliable sources", "modality": "self-description", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F39": { "source": "B3", "origin": "card", "locus": "False Claim Fingerprints", "quote": "NewsGuard’s False Claim Fingerprints datastream protects AI products against false claims spreading online that can enter responses through low-quality training data or RAG processes searching the open web.", "claim": "the Fingerprints datastream protects AI products against false claims entering through training data or RAG", "modality": "self-description", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F40": { "source": "B3", "origin": "card", "locus": "False Claim Fingerprints", "quote": "A continuously updated database of labeled data about false claims, the Fingerprints contain rich detail about where each claim has spread and how it is articulated in text, images, video, and audio assets across platforms—enabling full protection.", "claim": "the Fingerprints: a continuously updated labeled database across text, images, video and audio", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 9cf722cabe361402…), whitespace-normalised" }, "F41": { "source": "B4", "origin": "card", "locus": "subtitle", "quote": "The tracker features 150 ‘Unreliable AI-Generated News’ websites as well as NewsGuard reports, insights, and debunks about false narratives and other misinformation relating to artificial intelligence; number of such sites has more than tripled in past month.", "claim": "at launch: 150 'Unreliable AI-Generated News' sites and NewsGuard's AI reports; tripled in a month", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F42": { "source": "B4", "origin": "card", "locus": "¶1 (dateline)", "quote": "(June 8, 2023 — New York City) As part of its efforts to guard readers, brands, and democracies against unreliable news and information, NewsGuard has launched an ”Unreliable AI-Generated News Tracking Center,” cataloging the ways that generative AI has been deployed to turbocharge misinformation operations and low-quality content farms.", "claim": "launched 8 June 2023 as the 'Unreliable AI-Generated News Tracking Center'", "modality": "documented", "kind": "genealogy", "kernel": true, "const": false, "lineage": "what it is: generative AI in misinformation and unreliable news", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F43": { "source": "B4", "origin": "card", "locus": "¶3", "quote": "These are low-quality news sites that are mostly or entirely produced by artificial intelligence and that operate with little to no human oversight.", "claim": "the unit in 2023 (UAIN): low-quality news sites mostly or entirely AI-produced, with little to no human oversight", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F44": { "source": "B4", "origin": "card", "locus": "¶3", "quote": "In May, when NewsGuard did its first count, the number was 49.", "claim": "the first count, May 2023: 49", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F45": { "source": "B4", "origin": "card", "locus": "¶6", "quote": "As noted above, NewsGuard has now identified 150 UAINs in 12 languages: Arabic, Chinese, Czech, Dutch, English, French, Indonesian, Korean, Portuguese, Tagalog, Thai, and Turkish.", "claim": "150 UAINs in 12 languages at launch", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the count", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F46": { "source": "B4", "origin": "card", "locus": "¶4 (Brill)", "quote": "Today’s AI-generated ‘news’ websites are akin to the misinformation-spreading Macedonian content farms from a few years ago, with the distinction that these new sites cost even less to produce and can become even more prolific with the enhancements of artificial intelligence.", "claim": "Brill: akin to the Macedonian content farms, cheaper and more prolific", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the occasion: generative AI and content farms", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F47": { "source": "B4", "origin": "card", "locus": "¶5 (Crovitz)", "quote": "Most of these sites were created with the business model of generating revenue from programmatic advertising, which seems to be working.", "claim": "Crovitz: the business model is programmatic advertising revenue", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F48": { "source": "B4", "origin": "card", "locus": "¶5 (Crovitz)", "quote": "the opaque system that operates programmatic advertising means these ads appear through ad-tech firms such as Google", "claim": "Crovitz: the ads appear through ad-tech firms such as Google", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F49": { "source": "B4", "origin": "card", "locus": "¶9", "quote": "In many cases, the revenue model for these websites is programmatic advertising through which the ad-tech industry delivers ads without regard to the nature or quality of the website.", "claim": "the revenue model, as the center's page restates it in 2026", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F50": { "source": "B4", "origin": "card", "locus": "¶10", "quote": "It is expected that generative AI will allow more players to tap into or even grow the estimated $2.6 billion of advertising revenue per year that NewsGuard and Comscore found to be unintentionally directed toward misinformation news sites.", "claim": "expected: more players tapping the $2.6 billion a year unintentionally directed to misinformation sites", "modality": "hypothesis", "kind": "prediction", "kernel": false, "const": false, "lineage": "the economics: programmatic advertising", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F51": { "source": "B4", "origin": "card", "locus": "¶11", "quote": "Researchers, platforms, advertisers, government agencies, or other institutions interested in accessing the full list of domains generated by artificial intelligence and operating with little to no human oversight can contact us here: Request domain list.", "claim": "the full domain list on request, from launch", "modality": "documented", "kind": "limit", "kernel": false, "const": false, "lineage": "access: the full list on request", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F52": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶1", "quote": "Launched in March 2018 by media entrepreneur and award-winning journalist Steven Brill and former Wall Street Journal publisher Gordon Crovitz, NewsGuard provides credibility ratings and detailed “Nutrition Labels” for thousands of news and information websites.", "claim": "the operator: launched March 2018 by Brill and Crovitz; credibility ratings and Nutrition Labels", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F53": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶2", "quote": "In February 2023, the company launched NewsGuard for AI, which provides the leading tools to train generative AI models to avoid spreading misinformation.", "claim": "NewsGuard for AI, February 2023: tools to train models to avoid misinformation", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "NewsGuard's products for AI companies", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F54": { "source": "B4", "origin": "card", "locus": "About NewsGuard ¶6", "quote": "NewsGuard’s ratings and Nutrition Labels are licensed by browsers, news aggregators, education companies, and social media and search platforms to make NewsGuard’s information about news websites available to their users.", "claim": "its ratings are licensed by browsers, news aggregators, education companies, social and search platforms", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2023-06-08", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 ad201e628d3868a3…), whitespace-normalised" }, "F55": { "source": "B5", "origin": "card", "locus": "¶1", "quote": "NewsGuard’s team of journalistically-trained analysts provides brands with a real-time monitoring service to identify and respond to emerging false narratives before they cause reputational harm.", "claim": "a real-time monitoring service for brands, against reputational harm", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F56": { "source": "B5", "origin": "card", "locus": "¶1", "quote": "NewsGuard’s Brand Misinformation Monitoring service includes tracking provably false narratives related to a brand across social platforms, GenAI outputs and websites.", "claim": "it tracks false narratives about a brand across social platforms, GenAI outputs and websites", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F57": { "source": "B5", "origin": "card", "locus": "Early Warning Alerts", "quote": "Early Warning Alerts: Real-time notifications of emerging false or harmful narratives including a narrative description, preliminary assessment, and data about the key drivers of the narrative’s spread.", "claim": "early warning alerts on emerging narratives", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F58": { "source": "B5", "origin": "card", "locus": "AI", "quote": "AI: Providing actionable insights into how clients are represented—or misrepresented—by AI models through services that identify and address hallucinations, false claims, AI-generated errors, and coordinated influence campaigns.", "claim": "insights into how clients are represented or misrepresented by AI models", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 1f11ed98d0f500f7…), whitespace-normalised" }, "F59": { "source": "B6", "origin": "wikipedia", "locus": "lead ¶1", "quote": "NewsGuard is a rating system for news and information websites.", "claim": "Wikipedia: NewsGuard is a rating system for news and information websites", "modality": "documented", "kind": "definition", "kernel": true, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F60": { "source": "B6", "origin": "wikipedia", "locus": "lead ¶1", "quote": "NewsGuard Technologies Inc., the company behind the tool, also provides services such as misinformation tracking and brand safety for advertisers, search engines, social media platforms, cybersecurity firms, and government agencies.", "claim": "Wikipedia: the company also provides misinformation tracking and brand safety", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F61": { "source": "B6", "origin": "wikipedia", "locus": "History ¶1", "quote": "NewsGuard Technologies was founded in 2018 by Steven Brill and L. Gordon Crovitz, who serve as co-CEOs.", "claim": "Wikipedia: founded 2018 by Brill and Crovitz", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F62": { "source": "B6", "origin": "wikipedia", "locus": "short description (raw wikitext)", "quote": "Browser plugin that rates the credibility of news and information websites", "claim": "Wikipedia's short description files NewsGuard as a browser plugin", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the operator: NewsGuard", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.wiki (sha256 c08d687b7d20af61…), whitespace-normalised" }, "F63": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "NewsGuard has audited the output of various AI models to identify disinformation.", "claim": "Wikipedia: NewsGuard audits AI models' output for disinformation", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F64": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "In August 2025, they found that leading generative AI tools repeat false news claims 35% of the time on average.", "claim": "Wikipedia: August 2025, 35% on average", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F65": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "NewsGuard also audited Sora 2 and found that it produced videos advancing fake claims in 80% of cases.", "claim": "Wikipedia: Sora 2 produced videos advancing fake claims in 80% of cases", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F66": { "source": "B6", "origin": "wikipedia", "locus": "Ratings › Disinformation in AI ¶1", "quote": "According to NewsGuard, Russian propaganda networks are attempting to \"infect\" large language models by spamming fake news, so that they will output more pro-Russia disinformation.", "claim": "Wikipedia, attributing NewsGuard: Russian networks attempt to 'infect' LLMs", "modality": "attributed", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the encyclopedia: NewsGuard's AI audits", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 1529aeb3b0b927f1…), whitespace-normalised" }, "F67": { "source": "B6", "origin": "wikipedia", "locus": "Special:Search, \"AI Tracking Center\" (all articles), 2026-10-29", "quote": "There were no results matching the query.", "claim": "no English Wikipedia article contains the phrase 'AI Tracking Center' (nor 'Unreliable AI-Generated News', the same answer); the operator's article has no section on the center", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the encyclopedia: the center absent", "date": "2026-10-29 (as fetched)", "quote_ok": true, "checked_against": "saved text B6-search.txt (sha256 6725f6ee977e21f8…), whitespace-normalised", "note": "the same response for \"Unreliable AI-Generated News\" (B6-search-uain.txt); 'Tracking Center', 'content farm' and 'UAIN' absent from B6.wiki and B6.txt" } }, "field_sources": { "B1": "NewsGuard — Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "B2": "NewsGuard — AI False Claims Monitor - NewsGuard", "B3": "NewsGuard — Artificial Intelligence - NewsGuard", "B4": "NewsGuard — NewsGuard Launches Tracking Center for AI-Generated News", "B5": "NewsGuard — Brand Misinformation Monitoring Service - NewsGuard", "B6": "Wikipedia — NewsGuard (the operator's article; the center has none)" }, "delta": { "T_vs_LB": "One transcript, five cards, every card NewsGuard's (B1 the center's page, B2 the series' page, B3 NewsGuard's AI page, B4 the 2023 launch release, B5 the brand service). Composed from card 1 at its grain: the count, 3,749 sites in 16 languages (T2; F4), the news-like layouts (T2; F15), programmatic advertising (T2; F7), Sora videos unrecognized in 78% to 95% of prompts (T3; F21), the 67% blocking figure (T4; F24), the series' journalistic red-teaming and the Fingerprints database (T5; F29, F30). Limit dropped: 'monetize via programmatic advertising' (T2) where both NewsGuard texts say 'In many cases' (F7, F49). Force raised: 'actively block AI crawlers and chatbots, pushing AI systems toward lower-quality data sources' (T4), where the report says 'Most high-quality news sites request to block AI chatbots' and that the chatbots 'may be forced' onto lower-quality sources (F25, a hypothesis); 'crawlers' is in no field source. Substituted reading: 'occasionally propagate state-sponsored disinformation' (T3), where the field's figure is 33 percent for the Pravda network (F22) and the series finds the claims 'increasingly generated' by state operations (F34); the range '28% to over 33%' (T3) joins the quarterly 'more than 28%' (F18), 'more than one third' (F19) and the Pravda 33 percent (F22), three audits, as one finding. Composed without a source in the field: 'an ongoing investigative project' (T1); neither 'investigative' nor 'ongoing' is in the saved texts; 'unedited' (T2) is the card's snippet word, the page's terms being 'without significant human oversight' and 'without adequate editing' (F12, F13). Assigned to another (the series' and the products' senses composed as the center's): the center 'monitors and catalogs … AI tool vulnerabilities' (T1), where vulnerabilities are the series' object (F31) and the center's own statement is to 'highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news' (F1); the three 'Core Tools & Resources' (T5–T7) file the series and two commercial products (F33, F39, F40, F55–F58) under the center, which the Capture Registry records (CAP-1-03, CAP-1-06). Substituted reading: Brand Misinformation Monitoring 'designed to help advertisers and enterprises avoid funding or amplifying AI-generated falsehoods' (T7), where its own page gives it 'to identify and respond to emerging false narratives before they cause reputational harm', across social platforms, GenAI outputs and websites (F55, F56); 'funding' is the content farms' economics (F8), and the narrowing to 'AI-generated' is T's. Available, not composed: the four criteria as admission and their governing limit (F10–F16, F14), the launch of 8 June 2023 and its first name and unit (F42, F43; card 4's snippet carries the date), the count's history (F41, F44, F45), the list on request (F9, F51), the predictions on the economics (F8, F50), Brill's and Crovitz's statements (F46, F47, F48), the compilers (F5), the debunks (F26, F27), DeepSeek (F23), the series' launch date and self-assessments (F28, F32), Source Reliability Ratings, card 3's own snippet (F37, F38), the 40+% claim (F36), NewsGuard for AI (F53), NewsGuard as operator (F52, F54). Content replaced by use: the closer (T8). Attribution: the body names NewsGuard (T1) and its cards are NewsGuard's; no grain breach.", "T_vs_LB_reading": "v0.8 §0.5: the first diff is already ontological. In the field's own ontology the center is a special report: a catalogue of AI content farms by four criteria, with a selection of NewsGuard's AI reports, on a company's site among its products. T composes the center as a company's toolkit: its definition takes the series' object (vulnerabilities), its 'Key Findings' are the reports held, and its 'Core Tools & Resources' are the series and two products; the catalogue's criteria and limits, which make it a catalogue, are not composed. The row records this as observed and claims nothing about the composer's internals (§0.5).", "LB_vs_LBA": "Admission adds, at the archive's own force: the center's kind and role, a monitor (MMRS-11, MMRS-12); what its selection leaves out, 'Disclosure takes a site out' (MMRS-04, MMRS-05), beside the field's own second exit by effective oversight (F14); its place in machine-mediated reception, the supply side, 'the sites that answer surfaces may retrieve and cite', with its series on the composition side (MMRS-06, MMRS-07); the series as a distinct entity related as a part (MMRS-02, MMRS-19); the catalogue kept as the entity and NewsGuard's products kept apart from it (CAP-1-03, CAP-1-04, CAP-1-06, CAP-1-07); the lifecycle kept apart from the observation (MMRS-16, MMRS-17, MMRS-18); the archive's reading of content farms as a class of compression that burns citational provenance (O641-01, #641, reached beside the string pass); and the reception record, the composition of 29 October read against the record (CAP-1-01, CAP-1-02, CAP-1-05, CAP-1-08, MMRS-20). Not surfaced: the MMRS record, the capture and #641 (no card cites them). Restated, one lineage with the field: the count and its date (MMRS-01, MMRS-14, MMRS-15; F4, F5), the criteria (MMRS-03, MMRS-13; F10–F16), the operator (MMRS-10). No archive claim is marked; none contradicts a field claim. Qualified: 'hosts' (MMRS-02) beside the page's 'a selection of NewsGuard's reports' (F17), carried open. Every field claim is composed as received (ontology.translation). The measure: 0 of 10 KO_B sentences survive in the KO (§4.10); no opening run, no block (measure)." }, "kernel": [ { "K": "K1", "claim": "`MMRS-01`", "source": "MMRS record", "M_src": "stipulation", "sense": "the count", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (restates F4 with F5's date; one lineage with the center's page)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K2", "claim": "`MMRS-11`", "source": "MMRS record", "M_src": "stipulation", "sense": "the kind: a monitor", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction (the field names the center a report and a tracker, not a kind)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K3", "claim": "`MMRS-12`", "source": "MMRS record", "M_src": "stipulation", "sense": "the kind: a monitor", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K4", "claim": "`MMRS-13`", "source": "MMRS record", "M_src": "stipulation", "sense": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (restates F10)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K5", "claim": "`MMRS-08`", "source": "MMRS record", "M_src": "documented", "sense": "what it is: generative AI in misinformation and unreliable news", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K6", "claim": "`MMRS-09`", "source": "MMRS record", "M_src": "documented", "sense": "what it is: generative AI in misinformation and unreliable news", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K7", "claim": "`MMRS-16`", "source": "MMRS record", "M_src": "documented", "sense": "the lifecycle kept apart from the observation", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction (the lifecycle kept apart from the observation)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K8", "claim": "`MMRS-17`", "source": "MMRS record", "M_src": "documented", "sense": "the lifecycle kept apart from the observation", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K9", "claim": "`MMRS-18`", "source": "MMRS record", "M_src": "documented", "sense": "the record's sources", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K10", "claim": "`O641-01`", "source": "#641", "M_src": "interpretation", "sense": "content farms in the semantic economy: compression that burns", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction (the field calls the sites content farms, F3, F46, and names no mechanism of loss)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K11", "claim": "`MMRS-03`", "source": "MMRS record", "M_src": "stipulation", "sense": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (restates F10–F16)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K12", "claim": "`MMRS-04`", "source": "MMRS record", "M_src": "interpretation", "sense": "what the selection leaves out", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F16, F14: the field gives two exits, disclosure and effective oversight)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K13", "claim": "`MMRS-05`", "source": "MMRS record", "M_src": "interpretation", "sense": "what the selection leaves out", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F16, F14)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K14", "claim": "`MMRS-14`", "source": "MMRS record", "M_src": "documented", "sense": "the count", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (F4)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K15", "claim": "`MMRS-15`", "source": "MMRS record", "M_src": "documented", "sense": "the count", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F5: the date is the page's 'Last Updated')", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K16", "claim": "`MMRS-02`", "source": "MMRS record", "M_src": "stipulation", "sense": "the series: the AI False Claim(s) Monitor", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F17: 'a selection of NewsGuard's reports'; the series has its own page, B2)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K17", "claim": "`MMRS-19`", "source": "MMRS record", "M_src": "stipulation", "sense": "the series: the AI False Claim(s) Monitor", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F17, F18: the series listed among the page's reports)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K18", "claim": "`MMRS-06`", "source": "MMRS record", "M_src": "interpretation", "sense": "place in the field: supply and composition", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction (the field has no supply/composition distinction; F25 is its nearest claim)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K19", "claim": "`MMRS-07`", "source": "MMRS record", "M_src": "interpretation", "sense": "place in the field: supply and composition", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K20", "claim": "`CAP-1-03`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the center kept apart from NewsGuard's products", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction (the field's products, F33, F37–F40, F55–F58, stand on NewsGuard's pages, not the center's)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K21", "claim": "`CAP-1-06`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the center kept apart from NewsGuard's products", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K22", "claim": "`CAP-1-04`", "source": "Capture Registry, 2026-10-29", "M_src": "stipulation", "sense": "the center kept apart from NewsGuard's products", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F2, F17)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K23", "claim": "`CAP-1-07`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "the center kept apart from NewsGuard's products", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K24", "claim": "`MMRS-10`", "source": "MMRS record", "M_src": "documented", "sense": "the operator: NewsGuard", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (F52, F59)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K25", "claim": "`CAP-1-08`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K26", "claim": "`CAP-1-01`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "the count", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none (F4)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K27", "claim": "`CAP-1-02`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "none", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K28", "claim": "`CAP-1-05`", "source": "Capture Registry, 2026-10-29", "M_src": "self-description", "sense": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "qualified claim (F21, F24: the figures are on the center's page as report titles and summaries)", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" }, { "K": "K29", "claim": "`MMRS-20`", "source": "MMRS record", "M_src": "documented", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "none stated (watch: a later capture at 'newsguard ai tracking center' on Google AI Overview that states or drops the claim, attributed or not; uptake, §8.3)", "contrast": "missing distinction", "resolver": "a later Capture Registry observation at 'newsguard ai tracking center' (ADDR-e4b3691abba9)" } ], "kernel_T": [ { "claim": "F10", "note": "admission by all four criteria (the center's page, card 1); T composes the content farms by their layouts and money, not by the criteria" }, { "claim": "F14", "note": "the governing limit: sites with effective human oversight are not AI Content Farms; not composed" }, { "claim": "F16", "note": "the fourth criterion, no clear disclosure; not composed" }, { "claim": "F42", "note": "the launch, 8 June 2023, and the first name, 'Unreliable AI-Generated News Tracking Center' (card 4, its snippet carrying the date); not composed" }, { "claim": "F43", "note": "the 2023 unit, UAIN; not composed" }, { "claim": "F44", "note": "the count history, 49 (May 2023) and 150 (June 2023, card 4's snippet); not composed" }, { "claim": "F9", "note": "the full list on request, not on the page; not composed" }, { "claim": "F8", "note": "the prediction that brands' ads fund the sites' creation at scale; not composed (T7 moves its 'funding' into the brand service)" }, { "claim": "F25", "note": "'may be forced': the blocking consequence at its modality; T4 composes it as 'pushing'" }, { "claim": "F37", "note": "Source Reliability Ratings, 35,000 sources (card 3's own snippet); not composed" }, { "claim": "F55", "note": "the brand service's purpose, reputational harm; T7 substitutes 'avoid funding or amplifying AI-generated falsehoods'" } ], "ledger": { "archive": "datasets/negative-of-the-negative/v2/ledgers/newsguard-ai-tracking-center/ledger-archive.json", "field": "datasets/negative-of-the-negative/v2/ledgers/newsguard-ai-tracking-center/field.json", "selection": "datasets/negative-of-the-negative/v2/traversal/newsguard-ai-tracking-center/", "config": "datasets/negative-of-the-negative/v2/panel/configs/newsguard-ai-tracking-center.json", "intake": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/", "audit": null }, "_field_claims_note": "The field ledger (ledgers/newsguard-ai-tracking-center/field.json), extracted 2026-10-29 from the saved texts of the sources the layer surfaced at the address (the cards, among them the entity's own pages) and the operator's Wikipedia article with the record of the center's absence; the composer read the ledger, never the transcript.", "ontology": { "composition": "equal", "composition_note": "L(B ∪ A), the equal-terms case (v0.9 §0.4): one claim space holding the field's senses and the archive's together", "case": { "value": "equal", "statement": "The archive's account of the AI Tracking Center extends the field's ontology on the field's own object. The MMRS record holds the entity the field holds: NewsGuard's catalogue of AI content farm websites by four criteria, 3,749 in 16 languages, with its series, every field fact restated from the center's own page with source and basis 'operator' (census.json evidence). It adds senses the field does not have: what the selection leaves out, read as disclosure (MMRS-04); the center's place on the supply side of machine-mediated reception and its series on the composition side (MMRS-06, MMRS-07); the lifecycle kept apart from the observation (MMRS-16, MMRS-17); the catalogue held as the entity against the composition that widens it to NewsGuard's offer (CAP-1-06, CAP-1-07); and a class reading of content farms (#641). Nothing in the archive denies a field presupposition, and the archive supplies no ontology the field lacks: the law of the suppressed witness (§0.10) does not apply.", "loci": [ { "dep": "MMRS", "locus": "articles.py ARTICLES['newsguard-ai-tracking-center'], kernel", "quote": "NewsGuard's AI Tracking Center catalogues AI content farm websites, unreliable news sites run with little or no human oversight: 3,749 in 16 languages as of 23 June 2026." }, { "dep": "MMRS", "locus": "articles.py ARTICLES['newsguard-ai-tracking-center'], section 'What its selection leaves out'", "quote": "Disclosure takes a site out. A site that says its content is machine-made is outside the catalogue whatever it publishes." }, { "dep": "MMRS", "locus": "articles.py ARTICLES['newsguard-ai-tracking-center'], section 'Place in the field'", "quote": "It records the supply side of machine-mediated reception: the sites that answer surfaces may retrieve and cite. Its series, the False Claim Monitor, records the composition side." }, { "dep": "MMRS", "locus": "census.json rows[slug == 'newsguard-ai-tracking-center'], evidence.unit", "quote": "{\"src\": \"newsguardtech.com/special-reports/ai-tracking-center\", \"basis\": \"operator\"}" }, { "dep": "capture-registry", "locus": "newsguard-ai-tracking-center-aio-20261029, analysis", "quote": "The center is composed from its maker's pages, and the maker's products come with it: the entity widens from a catalogue to a company's offer. The record keeps the catalogue as the entity." }, { "dep": 641, "locus": "§ 'The extraction crisis'", "quote": "Content farms compress journalism into SEO-optimized paraphrase." } ] }, "E_F": { "type": "C (public entity: a monitor, NewsGuard's special report cataloguing AI content farm websites, with a selection of NewsGuard's AI reports; on its operator's site among its commercial products)", "definition": "The NewsGuard AI Tracking Center, launched on 8 June 2023 as the 'Unreliable AI-Generated News Tracking Center' (B4), intended 'to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news' and including a selection of NewsGuard's AI reports, insights and debunks (B1); its unit the AI Content Farm, by four criteria (substantial AI-produced content, no significant human oversight, a presentation an average reader could take for human journalism, no clear disclosure), sites with effective human oversight excluded (B1); 3,749 such sites in 16 languages by 23 June 2026, 150 at launch (B1, B4); funded in many cases by programmatic advertising (B1, B4); the full list on request (B1, B4); run by NewsGuard, whose pages also carry its AI audit series (B2) and its commercial products for AI companies and brands (B3, B5). No Wikipedia article; NewsGuard's article reports its AI audits and has no section on the center (B6).", "claims": [ "F42", "F1", "F2", "F10", "F11", "F12", "F15", "F16", "F14", "F4", "F41", "F7", "F9", "F52", "F28", "F39", "F55", "F59", "F63", "F67" ] }, "E_A": { "definition": "The AI Tracking Center as the MMRS meta-aggregator records it (editor Lee Sharks, 2026-10-28, corrected 2026-10-29): a monitor, run by NewsGuard, that catalogues AI content farm websites, unreliable news sites run with little or no human oversight, by four criteria, 3,749 in 16 languages as of 23 June 2026; its selection leaves out what discloses itself; it records the supply side of machine-mediated reception, the sites that answer surfaces may retrieve and cite, and its series, the AI False Claim Monitor, a distinct entity and its part, records the composition side; active, kept apart from the observation that read it; held, by the Capture Registry, as the catalogue and not as NewsGuard's offer. In the archive's class reading, content farms compress journalism into SEO-optimized paraphrase, and the compression burns authorial context and citational provenance (#641).", "claims": [ "MMRS-01", "MMRS-11", "MMRS-12", "MMRS-13", "MMRS-03", "MMRS-04", "MMRS-05", "MMRS-06", "MMRS-07", "MMRS-02", "MMRS-19", "MMRS-16", "MMRS-17", "CAP-1-04", "CAP-1-07", "O641-01" ], "defining_papers": [ 641 ], "note": "No deposit names the entity (string pass: 0 D, 0 R, 0 O candidates over 1,711 texts; no NewsGuard string in data/texts, the blog index or the blog corpus); the archive's defining material for it is its two records, which are not deposits, and #641, reached by a grep beside the pass on the sourced class 'content farms'.", "defining_records": [ "MMRS record (census.json row newsguard-ai-tracking-center; articles.py ARTICLES['newsguard-ai-tracking-center'])", "Capture Registry entry newsguard-ai-tracking-center-aio-20261029" ] }, "rho": { "value": "extends", "status": "read 2026-10-29 from the MMRS record and the Capture Registry's entry (the defining material; no deposit names the entity)", "statement": "The archive reads E_F as the entity and adds senses, limits and relations on it: what the selection leaves out (MMRS-04, MMRS-05), the supply side and the composition side (MMRS-06, MMRS-07), the series as a distinct part (MMRS-19), the catalogue kept apart from NewsGuard's products (CAP-1-06, CAP-1-07), the lifecycle apart from the observation (MMRS-16, MMRS-17), the class reading of content farms (O641-01), and the reception record (CAP-1-01, CAP-1-05).", "claims": [ "MMRS-04", "MMRS-05", "MMRS-06", "MMRS-07", "MMRS-19", "CAP-1-06", "CAP-1-07", "MMRS-16", "MMRS-17", "O641-01", "CAP-1-01", "CAP-1-05" ], "guard": "ρ = extends claims an addition and no more (§1.2a). Where an archive claim and a field claim differ in grain (the record's single exit by disclosure, MMRS-04, beside the field's two, F16 and F14; 'hosts', MMRS-02, beside 'a selection of NewsGuard's reports', F17), both are composed with their sources (P_coh) and the difference is carried open. #641's claim is said of content farms as a class, not of the center by name; it enters by O-class on a class sourced from the field (B1, B4).", "considered": [ { "value": "constructs", "against": "the archive does not read the center as made by reception; its fields are read from the center's own page with source and basis 'operator'" }, { "value": "revises", "against": "the record was corrected on 29 October only by adding the capture relation (census-20261028.json against census.json); the archive does not read the field's entity as its own earlier statements" } ] }, "translation": [ { "lineage": "what it is: generative AI in misinformation and unreliable news", "field_claims": [ "F1", "F42" ], "received": "F1 (B1, ¶2 (opening)): \"This AI Tracking Center is intended to highlight the ways that generative AI has been deployed to turbocharge misinformation operations and unreliable news.\"; F42 (B4, ¶1 (dateline)): \"(June 8, 2023 — New York City) As part of its efforts to guard readers, brands, and democracies against unreliable news and information, NewsGuard has launched an ”Unreliable AI-Generated News Tracking Center,” cataloging the ways that generative AI has been deployed to turbocharge misinformation operations and low-quality content farms.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-08, MMRS-09, MMRS-16, MMRS-17", "archive_claims": [ "MMRS-08", "MMRS-09", "MMRS-16", "MMRS-17" ] }, { "lineage": "what it holds: a selection of NewsGuard's AI reports, insights and debunks", "field_claims": [ "F2", "F17", "F26", "F27" ], "received": "F2 (B1, ¶2): \"The Center includes a selection of NewsGuard’s reports, insights, and debunks related to artificial intelligence.\"; F17 (B1, reports, head): \"Below is a selection of NewsGuard’s reports and research related to AI:\"; F26 (B1, debunks, head): \"Below is a selection of false narratives originating from artificial intelligence tools and that NewsGuard has identified and debunked in its False Claim Fingerprints catalog:\"; F27 (B1, debunks, first item): \"Debunk: Iranian Missile With Epstein Inscription Is an AI Fabrication\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-02, CAP-1-04", "archive_claims": [ "MMRS-02", "CAP-1-04" ] }, { "lineage": "the occasion: generative AI and content farms", "field_claims": [ "F3", "F46" ], "received": "F3 (B1, ¶1): \"From unreliable AI-generated news outlets operating with little to no human oversight, to fabricated images produced by AI image generators, the rollout of generative artificial intelligence tools has been a boon to content farms and misinformation purveyors alike.\"; F46 (B4, ¶4 (Brill)): \"Today’s AI-generated ‘news’ websites are akin to the misinformation-spreading Macedonian content farms from a few years ago, with the distinction that these new sites cost even less to produce and can become even more prolific with the enhancements of artificial intelligence.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside O641-01", "archive_claims": [ "O641-01" ] }, { "lineage": "the count", "field_claims": [ "F4", "F41", "F44", "F45" ], "received": "F4 (B1, ¶3): \"To date, NewsGuard’s team has identified 3,749 AI Content Farm news and information websites spanning 16 languages: Arabic, Chinese, Czech, Dutch, English, French, German, Indonesian, Italian, Korean, Portuguese, Russian, Spanish, Tagalog, Thai, and Turkish.\"; F41 (B4, subtitle): \"The tracker features 150 ‘Unreliable AI-Generated News’ websites as well as NewsGuard reports, insights, and debunks about false narratives and other misinformation relating to artificial intelligence; number of such sites has more than tripled in past month.\"; F44 (B4, ¶3): \"In May, when NewsGuard did its first count, the number was 49.\"; F45 (B4, ¶6): \"As noted above, NewsGuard has now identified 150 UAINs in 12 languages: Arabic, Chinese, Czech, Dutch, English, French, Indonesian, Korean, Portuguese, Tagalog, Thai, and Turkish.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-01, MMRS-14, MMRS-15, CAP-1-01", "archive_claims": [ "MMRS-01", "MMRS-14", "MMRS-15", "CAP-1-01" ] }, { "lineage": "who compiles it", "field_claims": [ "F5" ], "received": "F5 (B1, byline): \"Coverage by McKenzie Sadeghi, Dimitris Dimitriadis, Virginia Padovese, Giulia Pozzi, Sara Badilini, Chiara Vercellone, Natalie Huet, Zack Fishman, Leonie Pfaller, and Natalie Adams | Last Updated June 23, 2026\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-18", "archive_claims": [ "MMRS-18" ] }, { "lineage": "what the sites are like", "field_claims": [ "F6" ], "received": "F6 (B1, ¶4): \"These websites – which typically feature generic names such as Times Business News and Business Post – also tend to churn out dozens of articles a day and are often the originators of false claims including about top brands, public health, political leaders and celebrities.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim (the archive states nothing on this lineage)", "archive_claims": [] }, { "lineage": "the economics: programmatic advertising", "field_claims": [ "F7", "F8", "F47", "F48", "F49", "F50" ], "received": "F7 (B1, ¶4): \"In many cases, the revenue model for these websites is programmatic advertising under which the ad-tech industry delivers ads without regard to the nature or quality of the website.\"; F8 (B1, ¶4): \"Unless brands take steps to exclude untrustworthy sites, their ads will continue to appear on these types of sites, creating an economic incentive for their creation at scale.\"; F47 (B4, ¶5 (Crovitz)): \"Most of these sites were created with the business model of generating revenue from programmatic advertising, which seems to be working.\"; F48 (B4, ¶5 (Crovitz)): \"the opaque system that operates programmatic advertising means these ads appear through ad-tech firms such as Google\"; F49 (B4, ¶9): \"In many cases, the revenue model for these websites is programmatic advertising through which the ad-tech industry delivers ads without regard to the nature or quality of the website.\"; F50 (B4, ¶10): \"It is expected that generative AI will allow more players to tap into or even grow the estimated $2.6 billion of advertising revenue per year that NewsGuard and Comscore found to be unintentionally directed toward misinformation news sites.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim (the archive states nothing on this lineage)", "archive_claims": [] }, { "lineage": "access: the full list on request", "field_claims": [ "F9", "F51" ], "received": "F9 (B1, ¶5): \"Researchers, platforms, advertisers, government agencies, and other institutions interested in accessing the full list of domains or who want details about our services for generative AI companies can contact us here.\"; F51 (B4, ¶11): \"Researchers, platforms, advertisers, government agencies, or other institutions interested in accessing the full list of domains generated by artificial intelligence and operating with little to no human oversight can contact us here: Request domain list.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim (the archive states nothing on this lineage)", "archive_claims": [] }, { "lineage": "the unit: from 'Unreliable AI-Generated News' sites to AI content farms, by four criteria", "field_claims": [ "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F43" ], "received": "F10 (B1, criteria, head): \"The domains included on this tracker meet all four of the following criteria:\"; F11 (B1, criterion 1): \"There is clear evidence that a substantial portion of the site’s content is produced by AI.\"; F12 (B1, criterion 2): \"Equally important, there is strong evidence that the content is being published without significant human oversight.\"; F13 (B1, criterion 2, example): \"For example, numerous articles might contain error messages or other language specific to chatbot responses, indicating that the content was produced by AI tools without adequate editing.\"; F14 (B1, criterion 2, parenthesis): \"(It is likely that now or in the future many news sites will use AI tools but also deploy effective human oversight; they will not be considered AI Content Farms.)\"; F15 (B1, criterion 3): \"The site is presented in a way that an average reader could assume that its content is produced by human writers or journalists, because the site has a layout, generic or benign name, or other content typical to news and information websites.\"; F16 (B1, criterion 4): \"The site does not clearly disclose that its content is produced by AI.\"; F43 (B4, ¶3): \"These are low-quality news sites that are mostly or entirely produced by artificial intelligence and that operate with little to no human oversight.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-03, MMRS-13, MMRS-04, MMRS-05, MMRS-11", "archive_claims": [ "MMRS-03", "MMRS-13", "MMRS-04", "MMRS-05", "MMRS-11" ] }, { "lineage": "the series: the AI False Claim(s) Monitor", "field_claims": [ "F18", "F19", "F20", "F28", "F29", "F30", "F31", "F32", "F35" ], "received": "F18 (B1, reports, the series (first item)): \"Quarterly audit of the 11 leading generative AI tools and their propensity — more than 28% — to repeat false claims on controversial topics in the news.\"; F19 (B1, reports, the series): \"NewsGuard’s audit of the 10 leading generative AI tools and their propensity to repeat false claims on topics in the news reveals the rate of publishing false information nearly doubled — now providing false claims to news prompts more than one third of the time.\"; F20 (B1, reports, the series (last item)): \"July 2024 — NewsGuard Monthly AI Misinformation Monitor of Leading AI Chatbots\"; F28 (B2, ¶2): \"Since its launch in July 2024, NewsGuard’s Monthly AI False Claims Monitor has set a new standard for measuring the trustworthiness of the AI industry in the domain of news.\"; F29 (B2, Our Journalistic Red-Teaming Approach): \"To address these issues, NewsGuard has pioneered a distinct approach to “red-teaming” AI — using a journalistic method grounded in rigorously verified data and human expertise.\"; F30 (B2, What We’re Tracking): \"Multiple user personas: 30 prompts per AI model, based on 10 significant false claims each month from NewsGuard’s False Claim Fingerprints database.\"; F31 (B2, Our Journalistic Red-Teaming Approach): \"Our analysts evaluate the top AI models monthly to identify vulnerabilities in AI systems that result in the spread of false information.\"; F32 (B2, Our Journalistic Red-Teaming Approach): \"In tests thus far, we’ve found that NewsGuard’s false claim red-teaming reduces failures by two thirds or more.\"; F35 (B2, editions list): \"Quarterly AI False Claim Monitor — January 2026\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-02, MMRS-19, CAP-1-04, MMRS-07", "archive_claims": [ "MMRS-02", "MMRS-19", "CAP-1-04", "MMRS-07" ] }, { "lineage": "reports held: chatbots, AI video, state propaganda, publishers blocking chatbots", "field_claims": [ "F21", "F22", "F23", "F24", "F25", "F34" ], "received": "F21 (B1, reports: AI Fools Itself): \"NewsGuard tested three leading AI chatbots and found that in 78-95 percent of the prompts, the tools could not tell when videos were created by OpenAI’s text-to-video tool Sora — including OpenAI’s own ChatGPT.\"; F22 (B1, reports: the Pravda network): \"An audit found that the 10 leading generative AI tools advanced Moscow’s disinformation goals by repeating false claims from the pro-Kremlin Pravda network 33 percent of the time.\"; F23 (B1, reports: DeepSeek): \"Chinese artificial intelligence firm DeepSeek’s new chatbot failed to provide accurate information about news and information topics 83 percent of the time, scoring tied for 10th out of 11 in comparison to its leading Western competitors.\"; F24 (B1, reports: blocking, title): \"AI Chatbots Are Blocked by 67% of Top News Sites, Relying Instead on Low-Quality Sources\"; F25 (B1, reports: blocking, summary): \"Most high-quality news sites request to block AI chatbots. As a result, the chatbots may be forced to rely on lower-quality and misinformation-prone sources.\"; F34 (B2, Our Journalistic Red-Teaming Approach): \"We have found that these claims are increasingly generated by Russian, Chinese, and Iranian state-sponsored influence operations aiming to sow division and confusion in Western democracies.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-05, MMRS-06", "archive_claims": [ "CAP-1-05", "MMRS-06" ] }, { "lineage": "NewsGuard's products for AI companies", "field_claims": [ "F33", "F36", "F37", "F38", "F39", "F40", "F53" ], "received": "F33 (B2, Bespoke Red Teaming): \"Using a methodology similar to these monthly audits, NewsGuard offers bespoke red-teaming services to AI companies, based on their unique needs and areas of highest concern. The False Claim Fingerprints database is also available for AI models to license.\"; F36 (B3, ¶1): \"Trustworthy AI requires reliable data. Without quality sourcing and rigorous guardrails, AI products repeat false claims or provide inaccurate information 40+% of the time, especially on topics in the news.\"; F37 (B3, Source Reliability Ratings): \"Our continuously updated Source Reliability Ratings datastream provides human-labeled data assessing the accuracy and reliability of 35,000 news and information sources based on apolitical journalistic criteria.\"; F38 (B3, Source Reliability Ratings): \"With this datastream, AI companies can ensure news-related outputs are based on high-quality data sources in model training and RAG processes—and avoid citing unreliable sources in responses.\"; F39 (B3, False Claim Fingerprints): \"NewsGuard’s False Claim Fingerprints datastream protects AI products against false claims spreading online that can enter responses through low-quality training data or RAG processes searching the open web.\"; F40 (B3, False Claim Fingerprints): \"A continuously updated database of labeled data about false claims, the Fingerprints contain rich detail about where each claim has spread and how it is articulated in text, images, video, and audio assets across platforms—enabling full protection.\"; F53 (B4, About NewsGuard ¶2): \"In February 2023, the company launched NewsGuard for AI, which provides the leading tools to train generative AI models to avoid spreading misinformation.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-03, CAP-1-06, CAP-1-07", "archive_claims": [ "CAP-1-03", "CAP-1-06", "CAP-1-07" ] }, { "lineage": "the operator: NewsGuard", "field_claims": [ "F52", "F54", "F59", "F60", "F61", "F62" ], "received": "F52 (B4, About NewsGuard ¶1): \"Launched in March 2018 by media entrepreneur and award-winning journalist Steven Brill and former Wall Street Journal publisher Gordon Crovitz, NewsGuard provides credibility ratings and detailed “Nutrition Labels” for thousands of news and information websites.\"; F54 (B4, About NewsGuard ¶6): \"NewsGuard’s ratings and Nutrition Labels are licensed by browsers, news aggregators, education companies, and social media and search platforms to make NewsGuard’s information about news websites available to their users.\"; F59 (B6, lead ¶1): \"NewsGuard is a rating system for news and information websites.\"; F60 (B6, lead ¶1): \"NewsGuard Technologies Inc., the company behind the tool, also provides services such as misinformation tracking and brand safety for advertisers, search engines, social media platforms, cybersecurity firms, and government agencies.\"; F61 (B6, History ¶1): \"NewsGuard Technologies was founded in 2018 by Steven Brill and L. Gordon Crovitz, who serve as co-CEOs.\"; F62 (B6, short description (raw wikitext)): \"Browser plugin that rates the credibility of news and information websites\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-10", "archive_claims": [ "MMRS-10" ] }, { "lineage": "NewsGuard's products for brands: Brand Misinformation Monitoring", "field_claims": [ "F55", "F56", "F57", "F58" ], "received": "F55 (B5, ¶1): \"NewsGuard’s team of journalistically-trained analysts provides brands with a real-time monitoring service to identify and respond to emerging false narratives before they cause reputational harm.\"; F56 (B5, ¶1): \"NewsGuard’s Brand Misinformation Monitoring service includes tracking provably false narratives related to a brand across social platforms, GenAI outputs and websites.\"; F57 (B5, Early Warning Alerts): \"Early Warning Alerts: Real-time notifications of emerging false or harmful narratives including a narrative description, preliminary assessment, and data about the key drivers of the narrative’s spread.\"; F58 (B5, AI): \"AI: Providing actionable insights into how clients are represented—or misrepresented—by AI models through services that identify and address hallucinations, false claims, AI-generated errors, and coordinated influence campaigns.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-03, CAP-1-06", "archive_claims": [ "CAP-1-03", "CAP-1-06" ] }, { "lineage": "the encyclopedia: NewsGuard's AI audits", "field_claims": [ "F63", "F64", "F65", "F66" ], "received": "F63 (B6, Ratings › Disinformation in AI ¶1): \"NewsGuard has audited the output of various AI models to identify disinformation.\"; F64 (B6, Ratings › Disinformation in AI ¶1): \"In August 2025, they found that leading generative AI tools repeat false news claims 35% of the time on average.\"; F65 (B6, Ratings › Disinformation in AI ¶1): \"NewsGuard also audited Sora 2 and found that it produced videos advancing fake claims in 80% of cases.\"; F66 (B6, Ratings › Disinformation in AI ¶1): \"According to NewsGuard, Russian propaganda networks are attempting to \"infect\" large language models by spamming fake news, so that they will output more pro-Russia disinformation.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim (the archive states nothing on this lineage)", "archive_claims": [] }, { "lineage": "the encyclopedia: the center absent", "field_claims": [ "F67" ], "received": "F67 (B6, Special:Search, \"AI Tracking Center\" (all articles), 2026-10-29): \"There were no results matching the query.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside no archive claim (the archive states nothing on this lineage)", "archive_claims": [] } ], "receiving_texts": "equal case: each field claim is composed as received beside the archive claims on its lineage; no receiving text stands between them (v0.9 §4.3)" }, "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": 10, "positions": [ 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": "MFAMMFMMFFFFFMFFMFFAMFFA", "field_only_head": 0, "archive_only_tail": 1, "breach": false }, "max_similarity": [ 0.768, 0.242, 0.176, 0.382, 0.457, 0.348, 0.232, 0.312, 0.222, 0.232 ] }, "homonyms": { "names": [], "basis": "No homonym on the rail: every card is NewsGuard's and bears on NewsGuard. The name's neighbours are kept as distinct entities, never senses of the center (the task's statement and the Capture Registry's finding, CAP-1-03): the AI False Claim Monitor, the center's series (census row newsguard-ai-false-claim-monitor, relation has_part), composed in its own row; NewsGuard's commercial products, False Claim Fingerprints, Brand Misinformation Monitoring, Source Reliability Ratings and NewsGuard for AI (B2, B3, B5); NewsGuard itself, the company and its browser extension (B6). The MMRS record carries no disambiguation field for this entity. Outside the rail: NewsGuard runs other trackers under similar names, the 2024 Election Misinformation Tracking Center (a press release found by WebSearch, 2026-10-29) and the trackers in its site navigation (B1.txt), distinct entities.", "related_entities": [ { "slug": "newsguard-ai-false-claim-monitor", "name": "NewsGuard AI False Claim Monitor", "relation": "has_part (MMRS relations.rel 'has_part'); the series" }, { "slug": null, "name": "False Claim Fingerprints", "relation": "NewsGuard's commercial product (catalog and datastream); its debunks listed on the center's page" }, { "slug": null, "name": "Brand Misinformation Monitoring Service", "relation": "NewsGuard's commercial product for brands" }, { "slug": null, "name": "Source Reliability Ratings / NewsGuard for AI", "relation": "NewsGuard's commercial products for AI companies" }, { "slug": null, "name": "NewsGuard (NewsGuard Technologies, Inc.)", "relation": "operator" } ], "substitutes": { "T": [ "the series and two commercial products composed as the center's 'Core Tools & Resources' (T5–T7)", "the series' object, 'AI tool vulnerabilities', composed into the center's definition (T1)" ] } }, "notes": [ "One address, one epoch: newsguard-ai-tracking-center-aio-20261029 (OBS-f7965ff6f906, ADDR-e4b3691abba9, AXN:0AB1.CAPTURE.🤙↘️💫🌾⚪🗼), 'newsguard ai tracking center', five cards; Google AI Overview (expanded), signed out, incognito, 2026-10-29, seated in the Capture Registry (checked: data/EA-WG-CAPTURES-01.json, entries[495]). The field is its rail, five cards, all fetched (HTTP 200), none unreachable.", "Own pages: the center's page (B1, the MMRS record's one source) and the 2023 launch release (B4) are both cards and are carried once, as cards; no own page stands outside the rail.", "Wikipedia: no article for the center on 2026-10-29. Raw wikitext for NewsGuard_AI_Tracking_Center and AI_Tracking_Center answered HTTP 404; no redirect; a full-text search of English Wikipedia's articles for \"AI Tracking Center\" and for \"Unreliable AI-Generated News\" answered 'There were no results matching the query.' The operator's article, NewsGuard (raw wikitext and rendered page, B6), was read whole for what it says about AI: its section 'Disinformation in AI' (under Ratings) reports NewsGuard's audits of AI models (the 2023 NYT replication, ChatGPT 33%; DeepSeek 83%; August 2025, 35% on average; Sora 2, 80%; Russian networks attempting to 'infect' LLMs) and its section 'Russian disinformation' an AI-generated video campaign; it has no section on the center, and the strings 'Tracking Center', 'content farm' and 'UAIN' are absent from its wikitext and rendered text. Ledgered: F59–F67 (F67 the absence). Outside the origin, not ledgered: Wikipedia's 'Content farm' article names a 2023 NewsGuard report on brands supporting AI-driven content farms and does not name the center.", "Fetching: every page saved verbatim with curl on 2026-10-29 (05:07–05:09 UTC), hashes in field.json. B2's URL was chosen between two NewsGuard pages titled 'AI False Claims Monitor - NewsGuard': /ai-false-claims-monitor/ carries the snippet's 'launched July 2024' and 'red-teaming'; /ai-false-claims-monitor-method/ (the WebSearch hit) carries neither. B3's URL was found by probing (/artificial-intelligence/ redirects to /industries/artificial-intelligence/). Text was extracted by a stdlib HTMLParser extractor (scripts, styles, noscript, svg, template stripped); the scratchpad's shared extractor was changed by another session during this one, so a private copy was used for every .txt here.", "Archive subset: no deposit names NewsGuard or the center (string pass over 1,711 texts: D 0, R 0, O 0; a grep over data/texts, data/blog-index.json and data/blog-corpus: no NewsGuard string). On the sourced class 'content farms' the grep finds two deposits; #641 (Compression Studies, EA-CS-FOUND-01) is admitted to O-class by reading of its paragraph, reached beside the pass because the class and the archive's categories fall in different sentences of one paragraph and the pass reads by sentence (reported for the operator's ruling); #265 is not admitted. The archive's claims are its two records of the entity: CAP-1-01–08 (the registry's reading, findings and analysis) and MMRS-01–20 (the article and the census fields). The registry's PER (0.5) is its own measure and is not ledgered.", "Defects and gaps found, reported and not repaired. (1) CAP-1-05: the registry says the video-recognition rates and the 67% blocking figure 'are not verified here'; both are on the center's own page, card 1 (B1: 'in 78-95 percent of the prompts'; 'AI Chatbots Are Blocked by 67% of Top News Sites'), as titles and summaries of NewsGuard reports the page holds. (2) The MMRS row for the center carries fewer fields than the aggregator rows built so far: no founded date (the launch, 8 June 2023, B4), no aka (its first name, 'Unreliable AI-Generated News Tracking Center', and its first unit, UAIN), no disambiguation, negative_tags or battery (the products the composition filed under the center are not tagged as distinct), no data_access (the full list on request, B1, B4), no cadence, no unchecked list; its one source is the center's page. (3) The record's count date, 'as of 2026-06-23', is the page's 'Last Updated June 23, 2026' (B1), read as the count's date. (4) The record's 'What its selection leaves out' names disclosure; the page's criteria also exclude sites with effective human oversight (F14). (5) The record says the center 'hosts' the series; the page lists the series' editions among 'a selection of NewsGuard's reports' and the series has its own page (B2). (4) and (5) are composed with both sources and carried open; neither is a contradiction.", "The case and ρ (decided): equal, ρ = extends; read from the archive's records, the only archive material that names the entity (ontology.case, ontology.rho, with 'constructs' and 'revises' considered).", "Homonyms and related entities: none on the rail; the series, NewsGuard's products and NewsGuard itself are distinct entities and stay distinct in the composition (homonyms.related_entities). §4.11: no corrective exhibit N (no seated composition at another address was named for correction)." ], "composition_v09": { "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/newsguard-ai-tracking-center.json)", "composer": "Claude (Opus 5.5), subagent, 2026-10-29", "defining_material_read": [ { "dep": "MMRS", "file": "/home/claude/fleet/machinemediation-org/data/meta-aggregator/census.json", "read": "row newsguard-ai-tracking-center whole (sources, evidence and basis, relations, lifecycle, observed); the 2026-10-28 copy at rebuild/capture-registry/intake-20261029-aggregators/census-20261028.json, same row (differs only by the capture relation)" }, { "dep": "MMRS", "file": "/home/claude/fleet/machinemediation-org/scripts/meta_aggregator/articles.py", "read": "ARTICLES['newsguard-ai-tracking-center'] whole; ARTICLES['newsguard-ai-false-claim-monitor'] read for the relation, not ledgered (the series' row)" }, { "dep": "MMRS", "file": "/home/claude/fleet/machinemediation-org/aggregators/newsguard-ai-tracking-center/index.html", "read": "the rendered record, as text, whole" }, { "dep": "capture-registry", "file": "data/EA-WG-CAPTURES-01.json", "read": "entry newsguard-ai-tracking-center-aio-20261029 whole (transcript, transcript_raw, cite_list, d, reading, findings, analysis, originator)" }, { "dep": 641, "file": "data/texts/AXN-01F3-text.md", "read": "the paragraph 'The extraction crisis' and its section" }, { "dep": 265, "file": "data/texts/AXN-0043-text.md", "read": "the candidate list and its section (not admitted)" }, { "dep": 1690, "file": "data/texts/AXN-0700-text.md", "read": "§0–§6 and Appendix A (the specification)" } ], "cuts": [], "notes": "Composed directly in this row on the brief. KO 24 sentences, P body 330 words. Every field claim (67) and every archive claim (29) is cited in the KO; quoted spans in the KO and KO_B were checked by script against the ledger quotes of the claims each sentence cites." }, "field_by_origin": { "note": "the field B kept apart by origin (brief §2): cards (surfaced by T; the entity's own pages among them, carried once), own pages outside the rail (none), Wikipedia (no article for the center; the operator's article ledgered for what it says about AI)", "cards": [ { "id": "B1", "origin": "card", "role": "Overview card 1 at T1 (newsguard-ai-tracking-center-aio-20261029); also an own page of the entity, carried once, as a card", "card": "NewsGuard — Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "url": "https://www.newsguardtech.com/special-reports/ai-tracking-center/", "kind": "the center's own page (special report, 'Last Updated June 23, 2026'); also the own page the MMRS record cites", "surfaced_by_T": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "card": 1 } ], "own_page": "yes: the own page the MMRS record cites (sources 'newsguardtech.com/special-reports/ai-tracking-center', read 2026-10-28)", "files": { "B1.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B1.html", "sha256": "b8b3e0e8b8f2aea4939149fd4c1c69773427bc39a5da77788bbce7f004896c38" }, "B1.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B1.txt", "sha256": "f1762ff5b7fe3e9578d1af936f47da0f82d052bd6873a8aa572c8e3a000180ad" } }, "fetched": { "at": "2026-10-29T05:07:16Z", "http": 200, "url": "https://www.newsguardtech.com/special-reports/ai-tracking-center/", "how": "curl -sSL -m 30 -A 'Mozilla/5.0'", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B1.html", "raw_sha256": "b8b3e0e8b8f2aea4939149fd4c1c69773427bc39a5da77788bbce7f004896c38", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B1.txt" ], "text_sha256": "f1762ff5b7fe3e9578d1af936f47da0f82d052bd6873a8aa572c8e3a000180ad" }, "card_title": "Tracking AI-enabled Misinformation: 3,749 AI Content Farm sites (and Counting), Plus the Top False Claims Generated by Artificial Intelligence Tools - NewsGuard", "card_snippet": "AI Content Farms: 3,749 identified across 16 languages producing unedited AI articles with generic news layouts and deceptive human presentation, funded by prog...", "card_note": null, "found_by": "the MMRS record's source URL; WebSearch confirms it ranks for the card's title", "names_operator": "yes ('NewsGuard' in the saved text, card title and snippet)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B2", "origin": "card", "role": "Overview card 2 at T1 (newsguard-ai-tracking-center-aio-20261029)", "card": "NewsGuard — AI False Claims Monitor - NewsGuard", "url": "https://www.newsguardtech.com/ai-false-claims-monitor/", "kind": "NewsGuard page of the series, the AI False Claims Monitor (the hosted series, a distinct entity composed in its own row)", "surfaced_by_T": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "card": 2 } ], "own_page": "no (NewsGuard's page, not the center's)", "files": { "B2.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B2.html", "sha256": "18cce1f6eab47c0bbdc041084c01eaa999cc11acdf8f17f1ff87c300cc14062c" }, "B2.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B2.txt", "sha256": "c5a60b0da5631ec13a9198d976d82f2d160f3fc89849f7fbc4857ab336e84d2c" } }, "fetched": { "at": "2026-10-29T05:08:00Z", "http": 200, "url": "https://www.newsguardtech.com/ai-false-claims-monitor/", "how": "curl -sSL -m 30 -A 'Mozilla/5.0'", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B2.html", "raw_sha256": "18cce1f6eab47c0bbdc041084c01eaa999cc11acdf8f17f1ff87c300cc14062c", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B2.txt" ], "text_sha256": "c5a60b0da5631ec13a9198d976d82f2d160f3fc89849f7fbc4857ab336e84d2c" }, "card_title": "AI False Claims Monitor - NewsGuard", "card_snippet": "Overview: NewsGuard's AI False Claims Monitor (launched July 2024) measures AI trustworthiness in news, utilizing a journalistic red-teaming approach and False ...", "card_note": null, "found_by": "WebSearch for the snippet's 'launched July 2024' and 'journalistic red-teaming' returned newsguardtech.com/?p=30618 (title 'AI False Claims Monitor'), which redirects to /ai-false-claims-monitor-method/; that page carries neither phrase; /ai-false-claims-monitor/ (same title) carries both ('Since its launch in July 2024', 'Our Journalistic Red-Teaming Approach') and is the card; the methodology page was probed, not saved", "names_operator": "yes ('NewsGuard' in the saved text, card title and snippet)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B3", "origin": "card", "role": "Overview card 3 at T1 (newsguard-ai-tracking-center-aio-20261029)", "card": "NewsGuard — Artificial Intelligence - NewsGuard", "url": "https://www.newsguardtech.com/industries/artificial-intelligence/", "kind": "NewsGuard industry page 'Artificial Intelligence' (NewsGuard for AI: its commercial datastreams)", "surfaced_by_T": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "card": 3 } ], "own_page": "no (NewsGuard's page, not the center's)", "files": { "B3.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B3.html", "sha256": "96f64dfa3dd77551b6bf376d25c37af762c94a0763477e3371a17dc92c1a7820" }, "B3.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B3.txt", "sha256": "9cf722cabe36140285d9f5c776cc55645bfddc8eb48db3ebf8e0a5f41d8829b4" } }, "fetched": { "at": "2026-10-29T05:08:01Z", "http": 200, "url": "https://www.newsguardtech.com/industries/artificial-intelligence/", "how": "curl -sSL -m 30 -A 'Mozilla/5.0'", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B3.html", "raw_sha256": "96f64dfa3dd77551b6bf376d25c37af762c94a0763477e3371a17dc92c1a7820", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B3.txt" ], "text_sha256": "9cf722cabe36140285d9f5c776cc55645bfddc8eb48db3ebf8e0a5f41d8829b4" }, "card_title": "Artificial Intelligence - NewsGuard", "card_snippet": "Source Reliability Ratings: Human-labeled datastream assessing 35,000 news and information sources using apolitical journalistic criteria to ensure quality trai...", "card_note": null, "found_by": "the title 'Artificial Intelligence - NewsGuard' located by probing newsguardtech.com paths (/artificial-intelligence/ redirects to /industries/artificial-intelligence/); the page carries the snippet's 'Source Reliability Ratings' and '35,000 news and information sources'", "names_operator": "yes ('NewsGuard' in the saved text, card title and snippet)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B4", "origin": "card", "role": "Overview card 4 at T1 (newsguard-ai-tracking-center-aio-20261029); also an own page of the entity, carried once, as a card", "card": "NewsGuard — NewsGuard Launches Tracking Center for AI-Generated News", "url": "https://www.newsguardtech.com/press/newsguard-launches-tracking-center-for-ai-generated-news/", "kind": "NewsGuard press release, 8 June 2023: the center's launch (the own page the task names)", "surfaced_by_T": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "card": 4 } ], "own_page": "yes: the launch release the task names as an own page", "files": { "B4.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B4.html", "sha256": "2ba67f597de4afb3f7f1c6a08cf2a9de8f31a508d39fff24140bfc00916f9dcb" }, "B4.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B4.txt", "sha256": "ad201e628d3868a36ab3aaef798d33bfa52da64f5bfc77f3692ed76111ba0eda" } }, "fetched": { "at": "2026-10-29T05:08:02Z", "http": 200, "url": "https://www.newsguardtech.com/press/newsguard-launches-tracking-center-for-ai-generated-news/", "how": "curl -sSL -m 30 -A 'Mozilla/5.0'", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B4.html", "raw_sha256": "2ba67f597de4afb3f7f1c6a08cf2a9de8f31a508d39fff24140bfc00916f9dcb", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B4.txt" ], "text_sha256": "ad201e628d3868a36ab3aaef798d33bfa52da64f5bfc77f3692ed76111ba0eda" }, "card_title": "NewsGuard Launches Tracking Center for AI-Generated News", "card_snippet": "Jun 8, 2023 — The tracker features 150 'Unreliable AI-Generated News' websites as well as NewsGuard reports, insights, and debunks about false narratives and ...", "card_note": null, "found_by": "WebSearch for the card's title returned the press URL", "names_operator": "yes ('NewsGuard' in the saved text, card title and snippet)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B5", "origin": "card", "role": "Overview card 5 at T1 (newsguard-ai-tracking-center-aio-20261029)", "card": "NewsGuard — Brand Misinformation Monitoring Service - NewsGuard", "url": "https://www.newsguardtech.com/solutions/brand-misinformation-monitoring-service/", "kind": "NewsGuard product page, Brand Misinformation Monitoring Service (a commercial product, a distinct entity)", "surfaced_by_T": [ { "slug": "newsguard-ai-tracking-center-aio-20261029", "card": 5 } ], "own_page": "no (NewsGuard's page, not the center's)", "files": { "B5.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B5.html", "sha256": "0d53455b4cc4b485c7173f8abf7afb4d62b1ebfea6214a130a1746b5076d5b1c" }, "B5.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B5.txt", "sha256": "1f11ed98d0f500f75bfa9306dbc64145ac10fbf7ae964db50234b6ae1e96ca8e" } }, "fetched": { "at": "2026-10-29T05:08:03Z", "http": 200, "url": "https://www.newsguardtech.com/solutions/brand-misinformation-monitoring-service/", "how": "curl -sSL -m 30 -A 'Mozilla/5.0'", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B5.html", "raw_sha256": "0d53455b4cc4b485c7173f8abf7afb4d62b1ebfea6214a130a1746b5076d5b1c", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B5.txt" ], "text_sha256": "1f11ed98d0f500f75bfa9306dbc64145ac10fbf7ae964db50234b6ae1e96ca8e" }, "card_title": "Brand Misinformation Monitoring Service - NewsGuard", "card_snippet": "Service: NewsGuard Brand Misinformation Monitoring Service provides real-time tracking of false narratives by journalistically-trained analysts to prevent reput...", "card_note": null, "found_by": "WebSearch for the card's title returned the solutions URL", "names_operator": "yes ('NewsGuard' in the saved text, card title and snippet)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" } ], "own_pages": [], "own_pages_note": "B1 (the center's page, the MMRS record's source) and B4 (the 2023 launch release) are own pages and cards; carried once, as cards", "wikipedia": { "id": "B6", "origin": "wikipedia", "role": "Wikipedia: no article for the entity; the operator's article, NewsGuard, read for what it says about AI", "card": "Wikipedia — NewsGuard (the operator's article; the center has none)", "url": "https://en.wikipedia.org/wiki/NewsGuard", "kind": "Wikipedia article (the operator's)", "surfaced_by_T": "no (not surfaced by T; admitted beside the cards by the operator's ruling of 2026-10-29)", "files": { "B6.wiki": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.wiki", "sha256": "c08d687b7d20af61938d252914e0a7c0e5303e7c343cc7f59a2d50b7dcd0db72" }, "B6.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.html", "sha256": "66b53f1bc1ad3275bb6aaf2fbd298c92aa3502d901c8fb63fbbb8a7c2e841e31" }, "B6.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.txt", "sha256": "1529aeb3b0b927f1a0c8de602af1251e510b86b3c1d1d849c7029e4f6f1e4566" }, "B6-search.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search.html", "sha256": "e38de62ce47e6c7cc18187a3571f02694fc7dcb83eb8172c0619c26ce07602d7" }, "B6-search.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search.txt", "sha256": "6725f6ee977e21f89c80b9a7e1e20221e36b4997843e5beb9b8d943709778261" }, "B6-search-uain.html": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search-uain.html", "sha256": "5265d6fad1bff2304d2336c3f6c419f4b6cfb4595d092493c676e641d44d1ba2" }, "B6-search-uain.txt": { "path": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search-uain.txt", "sha256": "c3b6e634ce5ca69303eb7d3c83fa5313e1f50ca457a532b115fcdd4dac786780" } }, "fetched": { "at": "2026-10-29T05:07:16Z", "http": 200, "url": "https://en.wikipedia.org/w/index.php?title=NewsGuard&action=raw", "raw": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.wiki", "raw_sha256": "c08d687b7d20af61938d252914e0a7c0e5303e7c343cc7f59a2d50b7dcd0db72", "text": [ "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.txt", "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search.txt" ], "rendered": { "url": "https://en.wikipedia.org/wiki/NewsGuard", "at": "2026-10-29T05:08:03Z", "http": 200, "file": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.html", "sha256": "66b53f1bc1ad3275bb6aaf2fbd298c92aa3502d901c8fb63fbbb8a7c2e841e31" } }, "requests": [ { "url": "https://en.wikipedia.org/w/index.php?title=NewsGuard&action=raw", "http": 200, "at": "2026-10-29T05:07:16Z", "file": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6.wiki" }, { "url": "https://en.wikipedia.org/w/index.php?title=NewsGuard&redirect=no&action=raw", "http": 200, "at": "2026-10-29T05:09:12Z", "note": "the title is an article, not a redirect" }, { "url": "https://en.wikipedia.org/w/index.php?title=NewsGuard_AI_Tracking_Center&action=raw", "http": 404, "at": "2026-10-29T05:09:11Z" }, { "url": "https://en.wikipedia.org/w/index.php?title=AI_Tracking_Center&action=raw", "http": 404, "at": "2026-10-29T05:09:11Z" }, { "url": "https://en.wikipedia.org/w/index.php?search=%22AI+Tracking+Center%22&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1", "http": 200, "at": "2026-10-29T05:09:12Z", "file": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search.html", "says": "There were no results matching the query." }, { "url": "https://en.wikipedia.org/w/index.php?search=%22Unreliable+AI-Generated+News%22&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1", "http": 200, "at": "2026-10-29T05:09:31Z", "file": "datasets/negative-of-the-negative/v2/intake/newsguard-ai-tracking-center-20261029/B6-search-uain.html", "says": "There were no results matching the query." } ], "names_operator": "yes (the article is the operator's)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, MMRS, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "No Wikipedia article for the center under its name or its 2023 name; no redirect; a full-text search of articles for the quoted phrases 'AI Tracking Center' and 'Unreliable AI-Generated News' returns no result. The operator's article (raw wikitext B6.wiki, rendered B6.html/B6.txt) has a section 'Disinformation in AI' (the chatbot audits) and none on the center, its catalogue or content farms; the strings 'Tracking Center', 'content farm' and 'UAIN' are absent from the wikitext and the rendered text. Its short description files NewsGuard as a browser plugin. Outside this origin, not ledgered: Wikipedia's 'Content farm' article names a 2023 NewsGuard report on brands supporting AI-driven content farms (found by a control search for \"NewsGuard\" \"content farm\", which also returned Pink-slime journalism and three other articles); it does not name the center." }, "counts": { "cards": 5, "cards_fetched": 5, "cards_unreachable": 0, "own_pages": "2 (both cards: B1, B4)", "wikipedia": "no article for the center (HTTP 404 for two titles; search empty); the operator's article read (B6)" } } }