Examples
Examples from published documents
- "A large frontier developer shall write, implement, comply with, and clearly and conspicuously publish on its internet website a frontier AI framework" (California SB 53, section 22757.12(a))
- Seoul signatories undertook to publish "a safety framework focused on severe risks by the upcoming AI Summit in France" (Frontier AI Safety Commitments, AI Seoul Summit 2024)
- "G42 will publish non-sensitive, up-to-date and active copies of the Framework." (G42 Frontier Safety Framework, section 5)
Counter-examples
Not to be confused with
- Publishing a per-model report is not publishing the framework: SB 53 treats the frontier AI framework (section 22757.12(a)) and the per-model transparency report (section 22757.12(c)) as separate duties.
Related terms
Related AI safety terms
Terms defined under this element that are not NIKOLAI elements themselves, each as the cited document uses it. Every term is also listed in the AI safety glossary.
- Framework notification
- Framework notification, under the EU GPAI Code of Practice, is a provider's duty to give the AI Office unredacted access to its Safety and Security Framework, and to each update, within five business days of the framework being confirmed.
- Source: EU GPAI Code of Practice, Safety and Security Chapter
- Clearly and conspicuously publish
- Clearly and conspicuously publish is the standard California SB 53 sets for posting a frontier AI framework or transparency report on a developer's own website. The statute does not define the phrase further.
- Source: California SB 53
Source of record
Where this definition comes from
California SB 53, 22757.12(a)
“A large frontier developer shall write, implement, comply with, and clearly and conspicuously publish on its internet website a frontier AI framework that applies to the large frontier developer's frontier models”
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53EU GPAI Code of Practice, Safety and Security Chapter, Commitment 1, Measure 1.4
“Signatories will provide the AI Office with (unredacted) access to their Framework, and updates thereof, within five business days of either being confirmed.”
https://ec.europa.eu/newsroom/dae/redirection/document/118119Frontier AI Safety Commitments (AI Seoul Summit 2024), Preamble
“undertake to develop and deploy their frontier AI models and systems responsibly, in accordance with the following voluntary commitments, and to demonstrate how they have achieved this by publishing a safety framework focused on severe risks by the upcoming AI Summit in France”
https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024
Crosswalk
How named organisations use this concept
Every row below is a shadow mapping. A shadow row is CASRAI's own reading of a published document. No lab, evaluator or regulator named on a shadow row has declared, endorsed, or been consulted on it. That changes only when an organisation files its own Mapping Declaration.
| Organisation | Their term, as published | Match & verification | Source |
|---|---|---|---|
| California SB 53Shadow mapping California SB 53 | “22757.12(a): "A large frontier developer shall write, implement, comply with, and clearly and conspicuously publish on its internet website a frontier AI framework that applies to the large frontier developer's frontier models and describes how the large frontier developer approaches all of the following", ten topics including "(2) Defining and assessing thresholds", "(5) Using third parties to assess the potential for catastrophic risks", "(7) Cybersecurity practices to secure unreleased model weights" and "(9) Instituting internal governance practices to ensure implementation of these processes." Defined as "documented technical and organizational protocols to manage, assess, and mitigate catastrophic risks" (22757.11(g)).” | exactEQ confidence: high | California SB 53 |
| New York (RAISE Act, S8828)Shadow mapping NY RAISE Act (S8828) | “s.1421(1): a large frontier developer shall "write, implement, comply with, and clearly and conspicuously publish on its internet website a frontier AI framework" addressing paragraphs (a)-(j), including (d) "reviewing assessments and adequacy of mitigations as part of the decision to deploy a frontier model or use it extensively internally".” Statutory duty from 1 Jan 2027, according to the text of the Act that CASRAI reviewed; aligned with SB 53. | exactEQ confidence: high | NY RAISE Act (S8828) |
| EUShadow mapping EU GPAI Code of Practice, Safety and Security Chapter | “Commitment 1: "Signatories commit to adopting a state-of-the-art Safety and Security Framework" and to "notifying the AI Office of their Framework (as specified in Measure 1.4)": "Signatories will provide the AI Office with (unredacted) access to their Framework, and updates thereof, within five business days of either being confirmed." Public version only conditionally: "If and insofar as necessary to assess and/or mitigate systemic risks, Signatories will publish (e.g. via their websites) a summarised version of their Framework and Model Report(s)" (Measure 10.2).” CL: mandatory delivery to the regulator, conditional public summary; the framework must be confirmed 'no later than two weeks before placing the model on the market' (Measure 1.1). | closeCL confidence: high | EU GPAI Code of Practice, Safety and Security Chapter |
| Seoul Summit signatoriesShadow mapping Frontier AI Safety Commitments (Seoul 2024) | “Signatories undertake "to demonstrate how they have achieved this by publishing a safety framework focused on severe risks by the upcoming AI Summit in France."” Voluntary; the deadline was the February 2025 Paris summit. | exactEQ confidence: medium | Frontier AI Safety Commitments (AI Seoul Summit 2024) |
| US Congress (H.R. 9925, FRONTIER Act, not enacted)Shadow mapping FRONTIER Act (H.R. 9925) | “Sec. 4(a): "a large frontier developer shall write, implement, comply with, and clearly and conspicuously publish on a publicly available website of such developer a frontier AI framework", within one year of enactment or 90 days after first qualifying.” Introduced bill, not enacted. | exactEQ confidence: medium | FRONTIER Act, 119th Congress (H.R. 9925, introduced, not enacted) |
| AnthropicShadow mapping Anthropic Advanced AI Framework (legislative proposal) | “"Develop and publish a safety framework. Require Covered Developers to develop, publish, and follow a safety framework that sets out how they evaluate risks and how they make decisions about AI development and deployment." Minimum contents include "Identify the corporate officer primarily accountable for the framework's implementation." and "Describe the developer's process for modifying the safety framework." (p.5)” A legislative proposal for government, not a company practice. | exactEQ confidence: medium | Anthropic Advanced AI Framework (legislative proposal) |
| G42Shadow mapping G42 Frontier Safety Framework | “"Public Disclosure: G42 will publish non-sensitive, up-to-date and active copies of the Framework. We will share more detailed information with the UAE Government and relevant policy stakeholders." (s.5)” | exactEQ confidence: medium | G42 Frontier Safety Framework |
| OpenAIShadow mapping OpenAI Frontier Governance Framework | “"Under California's Transparency in Frontier AI Act (TFAIA), this FGF is our Frontier AI Framework, documenting OpenAI's technical and organizational protocols to manage, assess, and mitigate catastrophic risks, as defined under the TFAIA." (§1)” CL: a published framework that states which statutory duty it fulfils; see also mapping-declaration. | closeCL confidence: high | OpenAI Frontier Governance Framework |
| AnthropicShadow mapping Anthropic RSP v3.4 | “"If we update the RSP, we will publicly share the updated version prior to or on its effective date and will record the differences from the prior draft in the Change Log. We will maintain the current version of the RSP on our website." (§4(8))” CL: voluntary self-publication. The RSP states that statutory requirements such as SB 53 are addressed "in separate compliance frameworks" (fn 1). | closeCL confidence: high | Anthropic Responsible Scaling Policy v3.4 |
What do these codes mean?
- exact
- The source term is equivalent to this element
- close
- The source term is close but not equivalent to this element
- broad
- The source term is broader than this element
- narrow
- The source term is narrower than this element
- none
- No mapping claim — used for false-friend and declared-but-undefined rows
- EQ
- Equivalent
- CL
- Close
- BR
- Source is broader than the element
- NR
- Source is narrower than the element
- FF
- False friend — same or similar label, different meaning
- DU
- Declared but undefined by the source
- UV
- Unverified
Divergence
Where sources materially disagree
California SB 53, New York's RAISE Act and the proposed FRONTIER Act require public posting on the developer's website; the EU Code of Practice requires unredacted delivery to the AI Office within five business days and a public summary only 'if and insofar as necessary'; the Seoul Commitments set a one-time voluntary publication deadline.
