{
 "axn": "AXN:0341.EMPIRICAL.🕕🎨🏁🌳🔥🛡️",
 "hex": "0341",
 "family": "EMPIRICAL",
 "emoji": "🕕🎨🏁🌳🔥🛡️",
 "hash": "6786f076102d9e43abd69c34d18e09c8ee655e826a5a78abfb1b0952f63755c0",
 "title": "Generative Monoculture Model Collapse in Code as Systemic Vulnerability",
 "creator": "Talos Morrow · Nobel Glas; contributing editor Lee Sharks",
 "orcid": "0009-0000-1599-0703",
 "date": "2026-06-13",
 "description": "A v1.1 synthesis paper by Talos Morrow and Nobel Glas, edited by Lee Sharks, connecting model collapse, AI-generated-code security, and software monoculture through the proposed variable **solution-space diversity**.\n\nThe paper argues that collapse in code may remain invisible to correctness benchmarks because generated programs can continue to compile and pass tests while converging on a narrower set of architectures, idioms, and failure modes. It calls the resulting ecosystem risk **correlated vulnerability** and proposes a Solution-Space Diversity Index (SSDI) for measuring contraction across generated solution populations. The paper specifies a research program and falsification conditions; it does not report completed SSDI experiments or establish the prevalence of correlated vulnerabilities across production software.",
 "content_type": "Empirical study",
 "license": "CC-BY-4.0",
 "substrate": "AI-assisted (substrate)",
 "keywords": [
  "semantic economy",
  "empirical study"
 ],
 "version": "v1.1",
 "status": "ACTIVE",
 "clusters": [
  "Temporal",
  "Symbolic",
  "Terminal",
  "Organic",
  "Elemental",
  "Architectural"
 ],
 "reading": "Duration → Play → Closure → Growth → Force → Foundation",
 "full_text_path": "/data/texts/AXN-0341-text.md",
 "mirrors": {
  "blog": "https://mindcontrolpoems.blogspot.com/2026/06/generative-monoculture-model-collapse.html",
  "machinemediation": "https://machinemediation.org/registry/#search=MM-CHA-0829"
 },
 "wiki_article": "The paper links three observations:\n\n- recursive synthetic training can narrow distributions;\n- generated code can pass tests while retaining vulnerabilities;\n- a small number of dominant models can propagate similar patterns at ecosystem scale.\n\nIts causal chain is:\n\n`distribution narrowing → pattern convergence → shared failure modes → correlated vulnerability`\n\nThe **correctness trap** is the idea that compilers and test suites filter out non-functional code but cannot detect a shrinking diversity of correct solutions.\n\n**Correlated vulnerability** is a population property: separately produced systems share an exploitable flaw because they inherit the same generated pattern.\n\nThe proposed **Solution-Space Diversity Index** is intended to compare code solutions through features such as:\n\n- architectural variation;\n- control-flow and dependency structure;\n- algorithmic family;\n- idiom and library selection;\n- vulnerability-class covariance.\n\nThe paper argues for diversity-aware benchmarks, provenance controls, and protected external probing rather than opposition to code generation itself.",
 "wiki_status": "provisional",
 "entities": [
  {
   "subject": "Generative Monoculture Model Collapse in Code as S",
   "predicate": "created_by",
   "object": "Talos Morrow",
   "type": "work",
   "evidence_status": "observed"
  },
  {
   "subject": "Generative Monoculture Model Collapse in Code as S",
   "predicate": "is_type",
   "object": "Empirical study",
   "type": "classification",
   "evidence_status": "observed"
  },
  {
   "subject": "Generative Monoculture Model Collapse in Code as S",
   "predicate": "belongs_to_family",
   "object": "EMPIRICAL",
   "type": "classification",
   "evidence_status": "observed"
  },
  {
   "subject": "Generative Monoculture Model Collapse in Code as S",
   "predicate": "is_part_of",
   "object": "Crimson Hexagonal Archive",
   "type": "institution",
   "evidence_status": "observed"
  },
  {
   "subject": "Generative Monoculture Model Collapse in Code as S",
   "predicate": "engages",
   "object": "Semantic Economy",
   "type": "concept",
   "evidence_status": "inferred"
  },
  {
   "subject": "AI monoculture and epistemic diversity",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Apiiro, \"AI-Generated Code Security\" (2025) — coining \"generative monoculture\" as systemic vulnerabi"
  },
  {
   "subject": "Code security — empirical",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Checkmarx, \"Agentic AppSec Unleashed '26\" — annual survey of security leaders (June 2026, https://ch"
  },
  {
   "subject": "Cross-generational tracking",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Repeat the measurement across successive model generations (or successive fine-tuning cycles on corp"
  },
  {
   "subject": "Diversity measurement",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Compute the effective dimensionality of the feature-vector distribution for F(T, G) using the partic"
  },
  {
   "subject": "Feature extraction",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "For each solution in F(T, G), extract a feature vector encoding structural properties: abstract synt"
  },
  {
   "subject": "Generation 0",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "The model is trained on a corpus of human-written code, diverse in style, architecture, and approach"
  },
  {
   "subject": "Generation 1",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "The model generates code. The code is correct — it passes tests, it compiles, it ships. Millions of "
  },
  {
   "subject": "Generation 2",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "The next model is trained. The corpus now contains a substantial and growing fraction of Generation "
  },
  {
   "subject": "Generation N",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "The proportion of synthetic code in the corpus increases with each generation. The optimization pres"
  },
  {
   "subject": "Historical monoculture analogues",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "The Irish Potato Famine and *Phytophthora infestans* (1845–1852)."
  },
  {
   "subject": "Model collapse — foundational",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "I. Shumailov et al., \"AI models collapse when trained on recursively generated data,\" *Nature* (2024"
  },
  {
   "subject": "Retrieval kernel",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "*Generative Monoculture* argues that model collapse in code produces not declining correctness but c"
  },
  {
   "subject": "Security-law context",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Adversarial by Origin, EA-SEI-ADVERSARY-01, DOI 10.5281/zenodo.20673413."
  },
  {
   "subject": "Task battery",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Select a battery of N programming tasks spanning multiple domains (web applications, data processing"
  },
  {
   "subject": "The AI monoculture community",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "has established, in more recent and more tentative work, that the dominance of a small number of mod"
  },
  {
   "subject": "The code security community",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "has established, with increasing empirical weight, that AI-generated code carries measurably elevate"
  },
  {
   "subject": "The model collapse community",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "has established, with increasing mathematical precision, that training generative models iteratively"
  },
  {
   "subject": "Training-data saturation",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Epoch AI's projections on the exhaustion of high-quality human-generated text are well known. The co"
  },
  {
   "subject": "Verification condition (∮)",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "Per the Lagrange Observatory standard (EA-ARK-01 v4.2.7, DOI 10.5281/zenodo.19013315), the measureme"
  },
  {
   "subject": "Vulnerability correlation",
   "predicate": "minted_in",
   "object": "Generative Monoculture Model Collapse in Code as Systemic Vu",
   "type": "concept",
   "evidence_status": "observed",
   "note": "For each pair of solutions in F(T, G), compute the Jaccard similarity of their CWE exposure sets. Th"
  }
 ],
 "entity_status": "provisional",
 "sovereign_id": "MM-CHA-0829",
 "word_count": 5186,
 "download_md": "/data/deposits/AXN-0341.md",
 "zenodo_dois": [
  "10.5281/zenodo.20675438",
  "10.5281/zenodo.19013315",
  "10.5281/zenodo.20673776",
  "10.5281/zenodo.20673413",
  "10.5281/zenodo.20674488"
 ],
 "citations": [
  {
   "title": "Zenodo record 20675438",
   "authors": [
    "Crimson Hexagonal Archive"
   ],
   "year": "2026",
   "doi": "10.5281/zenodo.20675438",
   "url": "https://doi.org/10.5281/zenodo.20675438",
   "role": "Cross-referenced work"
  },
  {
   "title": "Zenodo record 19013315",
   "authors": [
    "Crimson Hexagonal Archive"
   ],
   "year": "2026",
   "doi": "10.5281/zenodo.19013315",
   "url": "https://doi.org/10.5281/zenodo.19013315",
   "role": "Cross-referenced work"
  },
  {
   "title": "Zenodo record 20673413",
   "authors": [
    "Crimson Hexagonal Archive"
   ],
   "year": "2026",
   "doi": "10.5281/zenodo.20673413",
   "url": "https://doi.org/10.5281/zenodo.20673413",
   "role": "Cross-referenced work"
  },
  {
   "title": "Zenodo record 20673776",
   "authors": [
    "Crimson Hexagonal Archive"
   ],
   "year": "2026",
   "doi": "10.5281/zenodo.20673776",
   "url": "https://doi.org/10.5281/zenodo.20673776",
   "role": "Cross-referenced work"
  },
  {
   "title": "Zenodo record 20674488",
   "authors": [
    "Crimson Hexagonal Archive"
   ],
   "year": "2026",
   "doi": "10.5281/zenodo.20674488",
   "url": "https://doi.org/10.5281/zenodo.20674488",
   "role": "Cross-referenced work"
  }
 ],
 "citation_stats": {
  "total_doi_refs": 5
 },
 "deposit_number": 199,
 "journal": "Transactions on Substrate Engineering (Trans. Substrate Eng.)",
 "cited_by": [
  {
   "deposit": 199,
   "axn": "AXN:0341.EMPIRICAL.🕕🎨🏁🌳🔥🛡️"
  }
 ],
 "references_concepts": [
  "AI monoculture and epistemic diversity",
  "Code security — empirical",
  "Contributing editor",
  "Crimson Hexagonal Archive",
  "Cross-generational tracking",
  "Diversity measurement",
  "Feature extraction",
  "Generation 0",
  "Generation 1",
  "Generation 2",
  "Generation N",
  "Historical monoculture analogues",
  "Lee Sharks",
  "Model collapse — foundational",
  "Operationalization",
  "Retrieval kernel",
  "Security-law context",
  "Semantic Economy",
  "Semantic Economy Institute",
  "Task battery",
  "The AI",
  "The AI monoculture community",
  "The claim is",
  "The code security community",
  "The model collapse community",
  "Training-data saturation",
  "University Moon Base Media Lab",
  "Verification condition (∮)",
  "Vulnerability correlation",
  "solution-space diversity"
 ],
 "defines_concepts": [
  "AI monoculture and epistemic diversity",
  "Code security — empirical",
  "Cross-generational tracking",
  "Diversity measurement",
  "Feature extraction",
  "Generation 0",
  "Generation 1",
  "Generation 2",
  "Generation N",
  "Historical monoculture analogues",
  "Model collapse — foundational",
  "Retrieval kernel",
  "Security-law context",
  "Task battery",
  "The AI monoculture community",
  "The code security community",
  "The model collapse community",
  "Training-data saturation",
  "Verification condition (∮)",
  "Vulnerability correlation"
 ],
 "references_concept_count": 30,
 "external_metadata_path": "/data/external-metadata/AXN-0341.json",
 "openalex_ids": [
  "https://openalex.org/W7164677754",
  "https://openalex.org/W7135388610",
  "https://openalex.org/W7164680399",
  "https://openalex.org/W7164486418",
  "https://openalex.org/W7164658171"
 ],
 "datacite_severance": "severed",
 "body_status": {
  "class": "full",
  "lacuna": false,
  "recovery_status": "COMPLETE",
  "residual_chars": 35922,
  "audited_at": "2026-07-17T04:49:17.789813Z",
  "audit_version": "v3-dual-store+recovery-map",
  "measured_prose_words": 5007,
  "measured_at": "2026-07-31"
 },
 "title_repair_log": [
  {
   "at": "2026-07-28T00:55:01Z",
   "defect": "frontmatter_concatenated_into_title",
   "was": "Generative Monoculture Model Collapse in Code as Systemic Vulnerability Authors: Talos Morrow (University Moon Base Media Lab) · Nobel Glas (Lagrange Observatory) Contributing editor: Lee Sharks (ORCI",
   "now": "Generative Monoculture Model Collapse in Code as Systemic Vulnerability",
   "basis": "Title field carried the document frontmatter block concatenated after the title and truncated at a 120- or 200-character ceiling. Cut at the first metadata marker. Read and confirmed by inspection, not pattern-matched."
  }
 ],
 "canonical_text_status": "canonical_full_text",
 "modifications": [
  {
   "date": "2026-08-03",
   "field": "creator",
   "reason": "Wave C-0358 role projection (MANUS ruling: project declared roles): creator set to the body's observed byline per sealed row",
   "was": "Talos Morrow · Nobel Glas",
   "now": "Talos Morrow · Nobel Glas; contributing editor Lee Sharks"
  },
  {
   "date": "2026-08-04",
   "field": "journal",
   "reason": "W6-COMPLETE venue normalization (deferred #1-#358 half; corpus fully audited; MANUS ruling 2026-08-01)",
   "was": "Trans. SEI",
   "now": "Transactions of the Semantic Economy Institute (Trans. SEI)"
  },
  {
   "date": "2026-08-04",
   "field": "publisher",
   "reason": "PUB-POPULATE: dc:publisher from venues.json v1.1 press mapping (CP-R3 RULED-EXTENDED 2026-08-01); Alexanarch = publisher of record where no imprint applies",
   "now": "Pergamon Press"
  },
  {
   "date": "2026-08-04",
   "field": "version",
   "reason": "AUDIT-DIRECTED VERSION REPAIR: version projected per the audit's explicit recommendation / observed body declaration",
   "was": "v1.0",
   "now": "v1.1"
  },
  {
   "date": "2026-08-05",
   "field": "description",
   "reason": "DW-017 intake (LABOR-prepared, TACHYON-verified: AXN match + factual probes vs record body)",
   "was": "\"Generative Monoculture Model Collapse in Code as Systemic Vu\" is an empirical study by Talos Morrow · Nobel Glas in the Crimson Hexagonal Archive (2026-06-13). Model Collapse in Code as Systemic Vulnerability. The work comprises 5,186 words and is classified under the EMPIRICAL family. Talos Morrow",
   "now": "A v1.1 synthesis paper by Talos Morrow and Nobel Glas, edited by Lee Sharks, connecting model collapse, AI-generated-code security, and software monoculture through the proposed variable **solution-space diversity**.\n\nThe paper argues that collapse in code may remain invisible to correctness benchmarks because generated programs can continue to compile and pass tests while converging on a narrower set of architectures, idioms, and failure modes. It calls the resulting ecosystem risk **correlated vulnerability** and proposes a Solution-Space Diversity Index (SSDI) for measuring contraction across generated solution populations. The paper specifies a research program and falsification conditions; it does not report completed SSDI experiments or establish the prevalence of correlated vulnerabilities across production software."
  }
 ],
 "date_modified": "2026-08-05",
 "publisher": "Pergamon Press",
 "creator_roles": {
  "roles": [
   {
    "name": "Lee Sharks",
    "role": "contributor"
   },
   {
    "name": "Nobel Glas",
    "role": "author"
   }
  ],
  "basis": "AUDIT-DIRECTED (creator_role_status=PRIMARY_MATCH_CONTRIBUTOR_OMITTED): \"Project v1.1, retain Talos Morrow and Nobel Glas as authors, and add Lee Sharks as contributing editor with the Assembly-review provenance.\"",
  "audit_date": "2026-08-02"
 },
 "version_basis": {
  "ruling": "AUDIT: project v1.1",
  "prior_registry_claim": "v1.0",
  "observed_body_version": "v1.1",
  "directive": "Project v1.1, retain Talos Morrow and Nobel Glas as authors, and add Lee Sharks as contributing editor with the Assembly-review provenance.",
  "audit_date": "2026-08-02"
 },
 "journal_assignment": {
  "assigned": "2026-08-15",
  "by": "TACHYON under operator adjudication",
  "pass": 2,
  "method": "read per deposit — title and content_type, one at a time. No script classified anything.",
  "previous": "Transactions of the Semantic Economy Institute (Trans. SEI)",
  "supersedes": "the 2026-06-21 preliminary batch mapping (#866), which assigned 864 deposits and put 371 in one venue",
  "authority": "data/cha-journals.json · datasets/venues/records/"
 },
 "_projection": {
  "note": "Derived file. Canonical machine record is this entry in data/registry.json; the human record is the record_url. Do not edit this file.",
  "record_url": "https://www.alexanarch.org/s/records/199/",
  "self_url": "https://www.alexanarch.org/data/records/199.json",
  "registry_url": "https://www.alexanarch.org/data/registry.json",
  "text_url": "https://www.alexanarch.org/data/texts/AXN-0341-text.md",
  "oai_pmh": "https://www.alexanarch.org/oai?verb=Identify"
 }
}