Source of record
Where this definition comes from
California SB 53, 22757.11(h), (j), (a)
“"Frontier developer": "a person who has trained, or initiated the training of, a frontier model..." (22757.11(h)); "Large frontier developer": "a frontier developer that together with its affiliates collectively had annual gross revenues in excess of five hundred million dollars ($500,000,000) in the preceding calendar year" (22757.11(j)).”
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53
Crosswalk
How named organisations use this concept
Every row below is a shadow mapping. It is CASRAI's own reading of a published document. No lab, evaluator or regulator named here has declared, endorsed, or been consulted on this mapping. That will change only when an organisation files its own Mapping Declaration — see the non-endorsement policy.
| Organisation | Their term, as published | Match | Source |
|---|---|---|---|
| Anthropic Advanced AI Framework / Risk Report | “"Covered Developer": developers that "Develop AI models requiring more than 10^25 training FLOP" and "Earn more than $500 million ... in annual AI-derived revenue, or spend more than $1 billion per year on AI research and development" (AAF p.3). The Risk Report has no developer field and describes its scope as "a risk assessment of Anthropic's activities as a whole" (§1).” Secondary citation {RR} (Anthropic Risk Report) confirms the report itself carries no developer field at all; the AAF definition is a separate, not-yet-enacted legislative proposal. | close confidence: medium | Anthropic Advanced AI Framework, legislative proposal |
| OpenAI Frontier Governance Framework | “Jurisdiction-specific entities: "OpenAI OpCo LLC is responsible for compliance with the TFAIA for covered models in the United States"; "OpenAI Ireland Limited is the provider of OpenAI's GPAI-SR models in the EU" (FGF §6, p.18).” | narrow confidence: high | OpenAI Frontier Governance Framework |
| xAI Frontier AI Framework / Grok 4.6 model card | “"xAI LLC ('xAI')" (FAIF s.1). The Grok 4.6 card adds: "SpaceXAI is a doing-business-as (DBA) name of XAI LLC. xAI and SpaceXAI may be used interchangeably throughout this card." (§1 fn, p.5).” This source's PDF metadata /Title reads "Privileged/Confidential DRAFT working FRAMEWORK DOC"; no xAI statement disambiguating draft vs. final status was found (open [VERIFY] item in the source document's register). | close confidence: medium | xAI Frontier AI Framework, 30 Jun 2026 |
| EU Safety and Security chapter | “"Providers of general-purpose AI models with systemic risk" can rely on the Safety and Security chapter (Commission landing page). The legal definition of "provider" itself is in the AI Act, which the source document did not fetch for this row.” Secondary citation tag {DEU} in the source document does not resolve to any URL in the provided sources table (distinct from the {DEV} discovery-sweep tag); treated here as unverified pending resolution. | close confidence: low | EU Commission GPAI Code of Practice landing page |
| California SB 53 SB 53 statute text | “"Frontier developer": "a person who has trained, or initiated the training of, a frontier model..." (22757.11(h)); "Large frontier developer": "a frontier developer that together with its affiliates collectively had annual gross revenues in excess of five hundred million dollars ($500,000,000) in the preceding calendar year" (22757.11(j)). Also "Affiliate" (22757.11(a)).” | exact confidence: high | California SB 53 |
| Demis Hassabis (personal essay) "A framework for frontier AI" essay | “"Organisations with 'Frontier Models' as defined by those benchmarks would be deemed 'Frontier Labs'" (§Framework ¶2).” Personal essay, not an official DeepMind policy document. | close confidence: medium | Demis Hassabis, "A framework for frontier AI and the dawning of a new age" |
| New York (RAISE Act, S8828) NY RAISE Act bill text | “Keeps the SB 53 developer tiers (para.).” | close confidence: medium | NY RAISE Act (S8828) |
| Amazon Amazon's Frontier Model Safety Framework | “No defined-term "developer" class; the Framework speaks in the first person ("Amazon") throughout and commits that "we will not deploy frontier AI models developed by Amazon that exceed specified risk thresholds without appropriate safeguards in place" (preamble, p.1).” The source document itself also lists Amazon among sources with "no developer term found" elsewhere in the same row's discussion (line 160: "Amazon (first-person only, no defined term)") -- a genuine internal inconsistency in the source, not an extraction error. The CL score reflects the first-person commitment language quoted in the bundled 'Other' row, not a formal defined term. | close confidence: low | Amazon's Frontier Model Safety Framework |
Gap
The SB 53 and FGF pattern (legal entity per jurisdiction) is the only one that supports attribution. The FGF also shows one developer holding different legal roles in different jurisdictions, so NIKOLAI needs `developer` x `jurisdiction` x `legal-role` rather than a single developer field.
