{
 "row": "mit-ai-risk-repository",
 "address": "mit ai risk repository",
 "entity": "The MIT AI Risk Repository (airisk.mit.edu/risks), the flagship of the MIT AI Risk Initiative (AIRI) at MIT FutureTech: a living database of risks extracted from published AI risk frameworks (version 4, December 2025: 1,700+ risks), with a Causal Taxonomy (entity, intent, timing) and a Domain Taxonomy (seven domains, 24 subdomains); its paper Slattery et al., arXiv:2408.12622 (v1 14 August 2024; v3 5 May 2026; Patterns 2026)",
 "row_type": "C (a public entity on which the archive bears: a database, one of the MMRS meta-aggregator's entities)",
 "aliases": [
  {"pattern": "\\bAI Risk Repository\\b", "flags": "i"},
  {"pattern": "\\bAI Risk Initiative\\b", "flags": "i"},
  {"pattern": "\\bAIRI\\b", "flags": ""},
  {"pattern": "\\bairisk\\.mit\\.edu\\b", "flags": "i"},
  {"pattern": "\\bMIT FutureTech\\b", "flags": "i"},
  {"pattern": "\\b2408\\.12622\\b", "flags": ""},
  {"pattern": "\\bSlattery\\b", "flags": ""}
 ],
 "classes": [
  {
   "term": "AI risk frameworks, taxonomies and databases",
   "pattern": "\\b(AI|artificial intelligence) risk (frameworks?|taxonom(y|ies)|databases?|repositor(y|ies)|classifications?)\\b",
   "flags": "i",
   "basis": "field: B6 (arXiv:2408.12622 v3, Abstract): 'We systematically analyzed every major AI risk framework published to date-74 frameworks containing 1,725 distinct risks'; B4 (FutureTech, v1 abstract): 'a systematic review of taxonomies and other structured classifications of AI risk'; B11 (airisk.mit.edu/risks, FAQ): 'What are some other databases of AI risks?'"
  }
 ],
 "classes_note": "One class, sourced from the field by locus: the repository as a database of AI risks built from AI risk frameworks and taxonomies, among other databases of AI risks (B6, B4, B11). The MMRS record's class (aggregators of machine-mediated reception work, 'roles: harm') is an archive source, ledgered as MMRS claims and not run as an O-class, since the brief confines classes to what B sources. Aliases: the entity's name without the MIT prefix (which matches the full name), its operator's name and abbreviation (AIRI, case-sensitive), its host, its research group, its paper's arXiv id and its lead author's surname. Admission by reading (§3.4): a candidate is admitted when its sentence or paragraph makes a claim bearing on the repository or on AI risk taxonomies as a class to which the archive applies its own category; a bare mention, a link list, a generic use of 'risk framework' about another object, or a homonym (another Slattery, another AIRI) is not a claim.",
 "exclude_deposits": [],
 "notes": "entity-keyed pass, 2026-10-29, on the brief rebuild/negative-of-the-negative/v09/BRIEF-NEW-ROW-AGGREGATOR.md; address 'mit ai risk repository' (mit-ai-risk-repository-aio-20261029, OBS-21a7cea64b70, ADDR-5762c6084031), Google AI Overview (expanded), signed out, incognito. The sibling mit-ai-incident-tracker (its own row) and the MIT AI Risk Navigator (a census candidate) are distinct entities."
}