{"entities":[{"id":"openai","name":"OpenAI"},{"id":"anthropic","name":"Anthropic"},{"id":"xai","name":"xAI"},{"id":"microsoft","name":"Microsoft"},{"id":"amazon","name":"Amazon"},{"id":"nvidia","name":"NVIDIA"},{"id":"rsp-2023","name":"Responsible Scaling Policy (2023)"},{"id":"google","name":"Google"},{"id":"rsp-v2-0","name":"Anthropic Responsible Scaling Policy — v2.0"},{"id":"rsp-v2-1","name":"Anthropic Responsible Scaling Policy — v2.1"},{"id":"rsp-v2-2","name":"Anthropic Responsible Scaling Policy — v2.2"},{"id":"rsp-v3-0","name":"Anthropic Responsible Scaling Policy — v3.0","shortName":"RSP · v3.0"},{"id":"rsp-v3-1","name":"Anthropic Responsible Scaling Policy — v3.1"},{"id":"rsp-v3-2","name":"Anthropic Responsible Scaling Policy — v3.2"},{"id":"rsp-v3-3","name":"Anthropic Responsible Scaling Policy — v3.3"},{"id":"rsp-v3-4","name":"Anthropic Responsible Scaling Policy — v3.4"},{"id":"preparedness-beta-2023","name":"OpenAI Preparedness Framework — Beta"},{"id":"frontier-governance-2026","name":"OpenAI Frontier Governance Framework — May 2026 edition","shortName":"FGF · May 2026 edition"},{"id":"faif-2025","name":"xAI Frontier Artificial Intelligence Framework — December 2025 edition","shortName":"FAIF · December 2025 edition"},{"id":"faif-2026","name":"xAI Frontier Artificial Intelligence Framework — June 2026 edition","shortName":"FAIF · June 2026 edition"},{"id":"meta","name":"Meta"},{"id":"cohere","name":"Cohere"},{"id":"zhipu","shortName":"Zhipu AI"},{"id":"meta-faf-v1-1","name":"Meta Frontier AI Framework — Version 1.1"},{"id":"ms-fgf-2025","name":"Microsoft Frontier Governance Framework — Version 1 (February 2025)"},{"id":"ms-fgf-2026","name":"Microsoft Frontier Governance Framework — February 2026"},{"id":"california-legislature","name":"California State Legislature"},{"id":"california-sb-53","name":"California SB 53 (2025): Transparency in Frontier Artificial Intelligence Act, as chaptered","shortName":"California SB 53","description":"Senate Bill 53 of the 2025–2026 Regular Session as chaptered (Chapter 138, Statutes of 2025). It adds Business and Professions Code §§ 22757.10–22757.16, Government Code § 11546.8 and Labor Code §§ 1107–1107.2. Its text describes obligations; it is not evidence of whom it applies to in practice or of compliance."},{"id":"anthropic-frontier-compliance-framework","name":"Anthropic Frontier Compliance Framework"},{"id":"us-white-house","name":"The White House"},{"id":"dsit","name":"Department for Science, Innovation and Technology","shortName":"DSIT"},{"id":"european-commission","name":"European Commission"},{"id":"inflection-ai","name":"Inflection AI"},{"id":"adobe","name":"Adobe"},{"id":"ibm","name":"IBM"},{"id":"palantir","name":"Palantir"},{"id":"salesforce","name":"Salesforce"},{"id":"scale-ai","name":"Scale AI"},{"id":"stability-ai","name":"Stability AI"},{"id":"g42","name":"G42"},{"id":"mistral-ai","name":"Mistral AI"},{"id":"naver","name":"Naver"},{"id":"samsung-electronics","name":"Samsung Electronics"},{"id":"technology-innovation-institute","name":"Technology Innovation Institute"},{"id":"magic-ai","name":"Magic"},{"id":"minimax","name":"MiniMax"},{"id":"zero-one-ai","name":"01.AI"},{"id":"wh-voluntary-ai-commitments-2023-07","name":"Ensuring Safe, Secure, and Trustworthy AI: voluntary AI commitments (July 2023)","shortName":"White House commitments (July 2023)","description":"The White House text of eight voluntary AI commitments under the headings safety, security and trust, announced on 21 July 2023 with seven companies. A commitment text, not evidence of what any company did."},{"id":"wh-voluntary-ai-commitments-2023-09","name":"Voluntary AI Commitments (September 2023)","shortName":"White House commitments (Sept. 2023)","description":"The White House text of the voluntary AI commitments linked from the fact sheet of 12 September 2023 on the second round of eight companies. Its scope clause differs from the July text."},{"id":"frontier-ai-safety-commitments-2024","name":"Frontier AI Safety Commitments, AI Seoul Summit 2024","shortName":"Seoul commitments (2024)","description":"Voluntary company commitments announced by the UK and Republic of Korea governments at the AI Seoul Summit on 21 May 2024 and published on GOV.UK, with a list of organisations that have agreed to them. A commitment text, not evidence of what any company did."},{"id":"gpai-code-of-practice-2025","name":"General-Purpose AI Code of Practice (final version, 10 July 2025)","shortName":"EU GPAI Code of Practice","description":"The EU General-Purpose AI Code of Practice, received and published by the European Commission on 10 July 2025. It consists of three separately authored chapters (Transparency, Copyright, and Safety and Security); provisions are quoted from each. Signing it is not evidence of compliance. A voluntary instrument under the EU AI Act."}],"claims":[{"id":"faif-2026-compliance-sentence","attributes":{"changeBasis":"document_comparison"}},{"id":"sb-calcompute-conditional-operation","statement":"SB 53's CalCompute section (Gov. Code § 11546.8, which establishes a consortium to develop a framework for a public cloud computing cluster) becomes operative only upon an appropriation in a budget act or other measure.","resolution":"supported","evidenceIds":["evidence-sb-calcompute-conditional-operation-1","evidence-sb-calcompute-conditional-operation-2"],"attributes":{"provision":{"kind":"exception","key":"sb53-calcompute-operative-condition"}}},{"id":"sb-frontier-model-threshold","statement":"SB 53 defines a 'frontier model' as a foundation model trained using more than 10^26 integer or floating-point operations. The count includes the original training run and any later fine-tuning, reinforcement learning or other material modifications that the developer applies to a preceding foundation model.","resolution":"supported","evidenceIds":["evidence-sb-frontier-model-threshold-1","evidence-sb-frontier-model-threshold-2"],"attributes":{"provision":{"kind":"threshold","key":"sb53-frontier-model-compute-threshold"}}},{"id":"sb-frontier-developer-definition","statement":"SB 53 defines a 'frontier developer' as a person who has trained, or initiated the training of, a frontier model, where that person has used or intends to use at least as much computing power to train it as the frontier-model definition specifies.","resolution":"supported","evidenceIds":["evidence-sb-frontier-developer-definition-1"],"attributes":{"provision":{"kind":"threshold","key":"sb53-frontier-developer-definition"}}},{"id":"sb-large-frontier-developer-threshold","statement":"SB 53 defines a 'large frontier developer' as a frontier developer that, together with its affiliates, had annual gross revenues above $500,000,000 in the preceding calendar year.","resolution":"supported","evidenceIds":["evidence-sb-large-frontier-developer-threshold-1","evidence-sb-large-frontier-developer-threshold-2"],"attributes":{"provision":{"kind":"threshold","key":"sb53-large-frontier-developer-revenue-threshold"}}},{"id":"sb-catastrophic-risk-threshold","statement":"SB 53 defines 'catastrophic risk' as a foreseeable and material risk that a frontier developer's development, storage, use or deployment of a frontier model will materially contribute to the death of, or serious injury to, more than 50 people, or to more than $1 billion in damage to, or loss of, property, arising from a single incident in which a frontier model (a) provides expert-level assistance in creating or releasing a chemical, biological, radiological or nuclear weapon; (b) engages without meaningful human oversight in a cyberattack or in conduct that would be murder, assault, extortion or theft if committed by a human; or (c) evades the control of its developer or user. Stated exclusions apply.","resolution":"supported","evidenceIds":["evidence-sb-catastrophic-risk-threshold-1","evidence-sb-catastrophic-risk-threshold-2","evidence-sb-catastrophic-risk-threshold-3"],"attributes":{"provision":{"kind":"threshold","key":"sb53-catastrophic-risk-definition"}}},{"id":"sb-framework-duty","statement":"SB 53 requires a large frontier developer to write, implement, comply with, and clearly and conspicuously publish on its website a frontier AI framework that applies to its frontier models. The framework must describe the developer's approach to ten listed matters.","resolution":"supported","evidenceIds":["evidence-sb-framework-duty-1","evidence-sb-framework-duty-2","evidence-sb-framework-duty-3"],"attributes":{"provision":{"kind":"required_action","key":"sb53-frontier-ai-framework-publication"}}},{"id":"sb-framework-annual-review","statement":"SB 53 requires a large frontier developer to review its frontier AI framework at least once a year and to update it as appropriate.","resolution":"supported","evidenceIds":["evidence-sb-framework-annual-review-1","evidence-sb-framework-annual-review-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-framework-annual-review"}}},{"id":"sb-framework-modification-30-days","statement":"SB 53 requires a large frontier developer that materially modifies its frontier AI framework to clearly and conspicuously publish the modified framework, with a justification for the modification, within 30 days.","resolution":"supported","evidenceIds":["evidence-sb-framework-modification-30-days-1","evidence-sb-framework-modification-30-days-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-framework-modification-publication"}}},{"id":"sb-transparency-report","statement":"SB 53 requires a frontier developer, before or at the same time as deploying a new or substantially modified frontier model, to publish clearly and conspicuously on its website a transparency report with listed basic information. A large frontier developer must also include summaries of its catastrophic-risk assessments under its frontier AI framework, their results, the extent of third-party evaluator involvement, and other steps taken to fulfil its framework for that model.","resolution":"supported","evidenceIds":["evidence-sb-transparency-report-1","evidence-sb-transparency-report-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-transparency-report-before-deployment"}}},{"id":"sb-system-card-deemed-compliance","statement":"SB 53 provides that a frontier developer that publishes the required transparency-report information as part of a larger document, such as a system card or model card, is deemed in compliance with the relevant paragraph.","resolution":"supported","evidenceIds":["evidence-sb-system-card-deemed-compliance-1"],"attributes":{"provision":{"kind":"exception","key":"sb53-transparency-report-larger-document"}}},{"id":"sb-internal-use-risk-summaries","statement":"SB 53 requires a large frontier developer to transmit to the Office of Emergency Services a summary of any assessment of catastrophic risk resulting from internal use of its frontier models. The summary is due every three months, or on another reasonable schedule that the developer specifies and communicates to OES in writing, with written updates as appropriate.","resolution":"supported","evidenceIds":["evidence-sb-internal-use-risk-summaries-1"],"attributes":{"provision":{"kind":"required_action","key":"sb53-internal-use-risk-summary-oes","parties":[{"role":"informed","actor":"Office of Emergency Services"}]}}},{"id":"sb-critical-incident-15-days","statement":"SB 53 requires a frontier developer to report any critical safety incident involving one or more of its frontier models to the Office of Emergency Services within 15 days of discovering it.","resolution":"supported","evidenceIds":["evidence-sb-critical-incident-15-days-1","evidence-sb-critical-incident-15-days-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-critical-safety-incident-report-15-days","parties":[{"role":"informed","actor":"Office of Emergency Services"}]}}},{"id":"sb-critical-incident-24-hours","statement":"SB 53 requires a frontier developer that discovers that a critical safety incident poses an imminent risk of death or serious physical injury to disclose that incident within 24 hours to an authority, including any law enforcement agency or public safety agency with jurisdiction, that is appropriate based on the nature of the incident and as required by law.","resolution":"supported","evidenceIds":["evidence-sb-critical-incident-24-hours-1"],"attributes":{"provision":{"kind":"required_action","key":"sb53-imminent-risk-disclosure-24-hours"}}},{"id":"sb-false-statement-prohibition","statement":"SB 53 prohibits a large frontier developer from making a materially false or misleading statement about its implementation of, or compliance with, its frontier AI framework. It also prohibits any frontier developer from making a materially false or misleading statement about catastrophic risk from its frontier models or about its management of catastrophic risk. The prohibition does not apply to a statement made in good faith that was reasonable under the circumstances.","resolution":"supported","evidenceIds":["evidence-sb-false-statement-prohibition-1","evidence-sb-false-statement-prohibition-2","evidence-sb-false-statement-prohibition-3"],"attributes":{"provision":{"kind":"required_action","key":"sb53-no-false-statements-framework-compliance"}}},{"id":"sb-federal-equivalence-pathway","statement":"SB 53 lets the Office of Emergency Services (OES) designate by regulation federal laws, regulations or guidance documents whose critical-safety-incident reporting standards are substantially equivalent to, or stricter than, those of Bus. & Prof. Code § 22757.13 and that are intended to assess, detect or mitigate catastrophic risk. A frontier developer that declares to OES its intent to comply through a designated federal standard is deemed in compliance with § 22757.13 to the extent it meets that standard, until it revokes the declaration or OES revokes the regulation; failing to meet the designated standard is a violation of the chapter.","resolution":"supported","evidenceIds":["evidence-sb-federal-equivalence-pathway-1","evidence-sb-federal-equivalence-pathway-2","evidence-sb-federal-equivalence-pathway-3","evidence-sb-federal-equivalence-pathway-4"],"attributes":{"provision":{"kind":"exception","key":"sb53-federal-incident-reporting-equivalence","parties":[{"role":"decides","actor":"Office of Emergency Services"}]}}},{"id":"sb-civil-penalty-cap","statement":"Under SB 53, a large frontier developer that fails to publish or transmit a required compliant document, makes a statement prohibited by § 22757.12(e), fails to report an incident as required, or fails to comply with its own frontier AI framework is subject to a civil penalty. The penalty depends on the severity of the violation and may not exceed $1,000,000 per violation.","resolution":"supported","evidenceIds":["evidence-sb-civil-penalty-cap-1"],"attributes":{"provision":{"kind":"threshold","key":"sb53-civil-penalty-cap"}}},{"id":"sb-attorney-general-enforcement","statement":"SB 53 provides that a civil penalty described in Bus. & Prof. Code § 22757.15 can be recovered only in a civil action brought by the California Attorney General.","resolution":"supported","evidenceIds":["evidence-sb-attorney-general-enforcement-1","evidence-sb-attorney-general-enforcement-2"],"attributes":{"provision":{"kind":"decision_authority","key":"sb53-attorney-general-enforcement","parties":[{"role":"decides","actor":"Attorney General"}]}}},{"id":"sb-whistleblower-anti-retaliation","statement":"SB 53 prohibits a frontier developer from making, adopting, enforcing or entering into a rule, regulation, policy or contract that prevents, or from retaliating against, a covered employee's disclosure of information to the Attorney General, a federal authority, a person with authority over the employee, or another covered employee with authority to investigate, discover or correct the issue, where the employee has reasonable cause to believe the information shows that the developer's activities pose a specific and substantial danger to public health or safety resulting from a catastrophic risk, or that the developer has violated the Act (Bus. & Prof. Code ch. 25.1).","resolution":"supported","evidenceIds":["evidence-sb-whistleblower-anti-retaliation-1","evidence-sb-whistleblower-anti-retaliation-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-covered-employee-anti-retaliation"}}},{"id":"sb-anonymous-internal-reporting","statement":"SB 53 requires a large frontier developer to provide a reasonable internal process through which a covered employee may anonymously disclose information the employee believes in good faith shows a specific and substantial danger to public health or safety resulting from a catastrophic risk, or a violation of the Act (Bus. & Prof. Code ch. 25.1). The process must give the discloser a monthly update on the investigation and actions taken, and disclosures and responses must be shared with the developer's officers and directors at least quarterly, except an officer or director whose wrongdoing is alleged.","resolution":"supported","evidenceIds":["evidence-sb-anonymous-internal-reporting-1","evidence-sb-anonymous-internal-reporting-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-anonymous-internal-disclosure-process","parties":[{"role":"informed","actor":"Officers and directors of the large frontier developer, except any officer or director whose wrongdoing is alleged in a disclosure or response"}]}}},{"id":"sb-enactment-2025-09-29","statement":"The Governor of California approved Senate Bill 53, which enacts the Transparency in Frontier Artificial Intelligence Act, on 29 September 2025; it was filed with the Secretary of State the same day and chaptered as Chapter 138, Statutes of 2025."},{"id":"sb-link-openai-fgf","statement":"OpenAI's Frontier Governance Framework (May 2026 edition) states that, under California's TFAIA, it is OpenAI's Frontier AI Framework.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-sb-link-openai-fgf-1","evidence-sb-link-openai-fgf-2"],"attributes":{"commitment":{"link":"designated_under"}}},{"id":"sb-link-rsp-v3-0","statement":"Anthropic's Responsible Scaling Policy v3.0 says it uses 'catastrophic risk' in its plain meaning rather than adopting any specific statutory definition. It adds that where laws such as California SB 53 define this or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-sb-link-rsp-v3-0-1","evidence-sb-link-rsp-v3-0-2","evidence-sb-link-rsp-v3-0-3"],"attributes":{"commitment":{"link":"names"}}},{"id":"sb-link-faif-2025","statement":"xAI's Frontier Artificial Intelligence Framework (December 2025 edition) states that the FAIF complies with California's Transparency in Frontier Artificial Intelligence Act (Bus. & Prof. Code § 22757.10 et seq.).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-sb-link-faif-2025-1"],"attributes":{"commitment":{"link":"complies_with"}}},{"id":"sb-link-ms-fgf-2026","statement":"Microsoft's Frontier Governance Framework (February 2026) explicitly references California's TFAIA in two places. Its change log describes the edition as a one-year update following finalisation of the EU GPAI Code of Practice, New York's RAISE Act and TFAIA. Its leading-indicator assessment is run on any model in scope for frontier-model requirements under applicable laws, and it names TFAIA among those laws.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-sb-link-ms-fgf-2026-1","evidence-sb-link-ms-fgf-2026-2"],"attributes":{"commitment":{"link":"responds_to"}}},{"id":"cm-wh-2023-07-published","statement":"On 21 July 2023 the White House published “Ensuring Safe, Secure, and Trustworthy AI”, a preamble and eight voluntary AI commitments under the headings safety, security and trust, with a fact sheet announcing that seven companies had chosen to undertake them immediately."},{"id":"cm-wh-2023-09-published","statement":"On 12 September 2023 the White House announced that it had secured a second round of voluntary commitments from eight companies and linked a commitments text headed “Voluntary AI Commitments”, with the same eight commitments and a scope clause measured against each company’s own most advanced model."},{"id":"cm-wh-2023-07-amazon","statement":"The White House named Amazon among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-amazon-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-anthropic","statement":"The White House named Anthropic among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-anthropic-1","evidence-cm-wh-2023-07-anthropic-2"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-google","statement":"The White House named Google among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-google-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-inflection-ai","statement":"The White House named Inflection AI among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-inflection-ai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-meta","statement":"The White House named Meta among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-meta-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-microsoft","statement":"The White House named Microsoft among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-microsoft-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-07-openai","statement":"The White House named OpenAI among the seven companies it convened on 21 July 2023, announcing that it had secured voluntary commitments from them to help move toward safe, secure and transparent development of AI technology.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-07-openai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-adobe","statement":"The White House named Adobe among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-adobe-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-cohere","statement":"The White House named Cohere among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-cohere-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-ibm","statement":"The White House named IBM among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-ibm-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-nvidia","statement":"The White House named Nvidia (NVIDIA) among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-nvidia-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-palantir","statement":"The White House named Palantir among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-palantir-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-salesforce","statement":"The White House named Salesforce among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-salesforce-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-scale-ai","statement":"The White House named Scale AI among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-scale-ai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-wh-2023-09-stability-ai","statement":"The White House named Stability AI (as “Stability”) among the eight companies from which it announced on 12 September 2023 that it had secured a second round of voluntary AI commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-wh-2023-09-stability-ai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"prov-wh-2023-07-scope","statement":"The July 2023 commitments text states that, where commitments mention particular models, they apply only to generative models that are overall more powerful than the current industry frontier, giving GPT-4, Claude 2, PaLM 2, Titan and, for image generation, DALL-E 2 as examples of released models.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-07-scope-1"],"attributes":{"provision":{"kind":"threshold","key":"scope"}}},{"id":"prov-wh-2023-07-duration","statement":"The July 2023 commitments text states that companies intend the voluntary commitments to remain in effect until regulations covering substantially the same issues come into force.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-07-duration-1"],"attributes":{"provision":{"kind":"exception","key":"duration"}}},{"id":"prov-wh-2023-07-red-teaming","statement":"Commitment 1 of the July 2023 text commits companies to internal and external red-teaming of models or systems in areas including misuse, societal risks and national security concerns, and to a red-teaming regime, drawing on independent domain experts, for all major public releases of new models within scope.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-07-red-teaming-1"],"attributes":{"provision":{"kind":"required_action","key":"red-teaming"}}},{"id":"prov-wh-2023-07-model-weights","statement":"Commitment 3 of the July 2023 text commits companies to treat unreleased model weights for models in scope as core intellectual property, with cybersecurity and insider-threat safeguards that include limiting access to model weights to those whose job function requires it.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-07-model-weights-1"],"attributes":{"provision":{"kind":"required_action","key":"model-weight-security"}}},{"id":"prov-wh-2023-07-public-reporting","statement":"Commitment 6 of the July 2023 text commits companies to publish reports for all new significant public model releases within scope, covering safety evaluations conducted, significant limitations in performance, effects on societal risks and the results of adversarial testing.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-07-public-reporting-1"],"attributes":{"provision":{"kind":"required_action","key":"public-reporting"}}},{"id":"prov-wh-2023-09-scope","statement":"The September 2023 commitments text states that, where commitments mention particular models, they apply only to generative models that are overall more powerful than the current most advanced model produced by the company making the commitment.","resolution":"supported","evidenceIds":["evidence-prov-wh-2023-09-scope-1"],"attributes":{"provision":{"kind":"threshold","key":"scope"}}},{"id":"provchange-wh-2023-09-scope","attributes":{"changeBasis":"document_comparison","provision":{"kind":"threshold","key":"scope"}}},{"id":"cm-link-rsp-2023-wh","statement":"Anthropic’s Responsible Scaling Policy v1.0 states that its ASL-2 deployment measures overlap substantially with the White House voluntary commitments that Anthropic and other companies made in July, which it says it continues to maintain.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-2023-wh-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"cm-link-rsp-v2-0-wh","statement":"Anthropic’s Responsible Scaling Policy v2.0 states that the policy also helps satisfy Anthropic’s Voluntary White House Commitments (2023).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-0-wh-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-link-rsp-v2-1-wh","statement":"Anthropic’s Responsible Scaling Policy v2.1 states that the policy also helps satisfy Anthropic’s Voluntary White House Commitments (2023).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-1-wh-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-link-rsp-v2-2-wh","statement":"Anthropic’s Responsible Scaling Policy v2.2 states that the policy also helps satisfy Anthropic’s Voluntary White House Commitments (2023).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-2-wh-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-seoul-published","statement":"The UK Department for Science, Innovation and Technology published the Frontier AI Safety Commitments on GOV.UK on 21 May 2024, the opening day of the AI Seoul Summit, and updated the page on 7 February 2025 to add organisations to its list."},{"id":"cm-seoul-amazon","statement":"Amazon is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-amazon-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-anthropic","statement":"Anthropic is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-anthropic-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-cohere","statement":"Cohere is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-cohere-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-google","statement":"Google is listed, as “Google”, among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-google-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-g42","statement":"G42 is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-g42-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-ibm","statement":"IBM is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-ibm-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-inflection-ai","statement":"Inflection AI is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-inflection-ai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-meta","statement":"Meta is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-meta-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-microsoft","statement":"Microsoft is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-microsoft-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-mistral-ai","statement":"Mistral AI is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-mistral-ai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-naver","statement":"Naver is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-naver-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-openai","statement":"OpenAI is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-openai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-samsung-electronics","statement":"Samsung Electronics is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-samsung-electronics-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-technology-innovation-institute","statement":"Technology Innovation Institute is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-technology-innovation-institute-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-xai","statement":"xAI is listed among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-xai-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-zhipu","statement":"Zhipu AI is listed, as “Zhipu.ai”, among the sixteen organisations that, according to the GOV.UK text of 21 May 2024, the UK and Republic of Korea governments announced had agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-zhipu-1"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-magic-ai","statement":"Magic is listed, as “Magic”, among four organisations added on 7 February 2025 to the GOV.UK list of organisations that have agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-magic-ai-1","evidence-cm-seoul-magic-ai-2"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-minimax","statement":"MiniMax is listed, as “Minimax”, among four organisations added on 7 February 2025 to the GOV.UK list of organisations that have agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-minimax-1","evidence-cm-seoul-minimax-2"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-zero-one-ai","statement":"01.AI is listed, as “01.ai”, among four organisations added on 7 February 2025 to the GOV.UK list of organisations that have agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-zero-one-ai-1","evidence-cm-seoul-zero-one-ai-2"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-seoul-nvidia","statement":"NVIDIA is listed, as “NVIDIA”, among four organisations added on 7 February 2025 to the GOV.UK list of organisations that have agreed to the Frontier AI Safety Commitments.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-seoul-nvidia-1","evidence-cm-seoul-nvidia-2"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"prov-fasc-2024-publish-framework","statement":"The Frontier AI Safety Commitments state that the listed organisations undertake to demonstrate how they develop and deploy frontier AI responsibly by publishing a safety framework focused on severe risks by the upcoming AI Summit in France.","resolution":"supported","evidenceIds":["evidence-prov-fasc-2024-publish-framework-1"],"attributes":{"provision":{"kind":"required_action","key":"publish-safety-framework"}}},{"id":"prov-fasc-2024-intolerable-thresholds","statement":"Commitment II states that the organisations will set out thresholds at which severe risks posed by a model or system, unless adequately mitigated, would be deemed intolerable.","resolution":"supported","evidenceIds":["evidence-prov-fasc-2024-intolerable-thresholds-1"],"attributes":{"provision":{"kind":"threshold","key":"intolerable-risk-thresholds"}}},{"id":"prov-fasc-2024-threshold-input","statement":"Commitment II states that the thresholds should be defined with input from trusted actors, including the organisations’ respective home governments as appropriate.","resolution":"supported","evidenceIds":["evidence-prov-fasc-2024-threshold-input-1"],"attributes":{"provision":{"kind":"decision_authority","key":"threshold-input","parties":[{"role":"consulted","actor":"Trusted actors, including home governments as appropriate"}]}}},{"id":"prov-fasc-2024-no-development","statement":"Commitment IV states that, in the extreme, the organisations commit not to develop or deploy a model or system at all if mitigations cannot be applied to keep risks below the thresholds.","resolution":"supported","evidenceIds":["evidence-prov-fasc-2024-no-development-1"],"attributes":{"provision":{"kind":"required_action","key":"no-development-above-thresholds"}}},{"id":"prov-fasc-2024-transparency-exception","statement":"Commitment VII states that the organisations will provide public transparency on the implementation of commitments I–VI, except insofar as doing so would increase risk or divulge sensitive commercial information to a degree disproportionate to the societal benefit.","resolution":"supported","evidenceIds":["evidence-prov-fasc-2024-transparency-exception-1"],"attributes":{"provision":{"kind":"exception","key":"transparency-exception"}}},{"id":"cm-link-rsp-v2-0-seoul","statement":"Anthropic’s Responsible Scaling Policy v2.0 states that the policy also helps satisfy Anthropic’s Frontier AI Safety Commitments (2024).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-0-seoul-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-link-rsp-v2-1-seoul","statement":"Anthropic’s Responsible Scaling Policy v2.1 states that the policy also helps satisfy Anthropic’s Frontier AI Safety Commitments (2024).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-1-seoul-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-link-rsp-v2-2-seoul","statement":"Anthropic’s Responsible Scaling Policy v2.2 states that the policy also helps satisfy Anthropic’s Frontier AI Safety Commitments (2024).","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-rsp-v2-2-seoul-1"],"attributes":{"commitment":{"scope":"helps satisfy","link":"fulfils"}}},{"id":"cm-link-ms-fgf-2025-seoul","statement":"Microsoft’s Frontier Governance Framework, Version 1 (February 2025), states that the framework has its genesis in the voluntary Frontier AI Safety Commitments that Microsoft and fifteen other AI labs made in May 2024.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-ms-fgf-2025-seoul-1"],"attributes":{"commitment":{"link":"responds_to"}}},{"id":"cm-link-ms-fgf-2026-seoul","statement":"Microsoft’s Frontier Governance Framework, February 2026 edition, states that the framework has its genesis in the voluntary Frontier AI Safety Commitments that Microsoft and fifteen other AI labs made in May 2024.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-ms-fgf-2026-seoul-1"],"attributes":{"commitment":{"link":"responds_to"}}},{"id":"cm-link-meta-faf-v1-1-seoul","statement":"Meta’s Frontier AI Framework v1.1 opens its scope section by stating that, in line with the Frontier AI Safety Commitments, which it says Meta signed in May 2024, the framework relates to Meta’s forthcoming most advanced models and systems.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-meta-faf-v1-1-seoul-1"],"attributes":{"commitment":{"link":"responds_to"}}},{"id":"cm-eu-code-published","statement":"The European Commission received and published the final General-Purpose AI Code of Practice on 10 July 2025. The Code consists of three separately authored chapters: Transparency, Copyright, and Safety and Security."},{"id":"cm-eu-code-amazon","statement":"Amazon appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-amazon-1","evidence-cm-eu-code-amazon-2","evidence-cm-eu-code-amazon-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-anthropic","statement":"Anthropic appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-anthropic-1","evidence-cm-eu-code-anthropic-2","evidence-cm-eu-code-anthropic-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-cohere","statement":"Cohere appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-cohere-1","evidence-cm-eu-code-cohere-2","evidence-cm-eu-code-cohere-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-google","statement":"Google appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-google-1","evidence-cm-eu-code-google-2","evidence-cm-eu-code-google-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-ibm","statement":"IBM appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-ibm-1","evidence-cm-eu-code-ibm-2","evidence-cm-eu-code-ibm-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-microsoft","statement":"Microsoft appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-microsoft-1","evidence-cm-eu-code-microsoft-2","evidence-cm-eu-code-microsoft-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-mistral-ai","statement":"Mistral AI appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-mistral-ai-1","evidence-cm-eu-code-mistral-ai-2","evidence-cm-eu-code-mistral-ai-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-openai","statement":"OpenAI appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-openai-1","evidence-cm-eu-code-openai-2","evidence-cm-eu-code-openai-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"cm-eu-code-xai","statement":"From the list’s first publication on 1 August 2025, the European Commission’s signatory page leaves xAI out of its list of signatories of the General-Purpose AI Code of Practice and states separately that xAI signed up to the Safety and Security chapter.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-xai-1"],"attributes":{"commitment":{"stage":"listed","scope":"Safety and Security chapter only"}}},{"id":"cm-eu-code-intent-openai","statement":"On 11 July 2025 OpenAI announced its intention to sign the General-Purpose AI Code of Practice, subject to the then-current version being formally approved by the AI Board during its adequacy assessment.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-intent-openai-1"],"attributes":{"commitment":{"stage":"intent_announced","scope":"conditional on formal approval by the AI Board"}}},{"id":"cm-eu-code-intent-anthropic","statement":"On 21 July 2025 Anthropic announced that, after review, it intended to sign the General-Purpose AI Code of Practice.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-intent-anthropic-1"],"attributes":{"commitment":{"stage":"intent_announced"}}},{"id":"cm-eu-code-intent-google","statement":"On 30 July 2025 Google announced, in a post by Kent Walker, that it would join other companies in signing the General-Purpose AI Code of Practice.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-intent-google-1"],"attributes":{"commitment":{"stage":"intent_announced"}}},{"id":"cm-eu-code-intent-xai","statement":"On 31 July 2025 xAI announced on X that it would sign the Safety and Security chapter of the EU AI Act’s Code of Practice.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-intent-xai-1"],"attributes":{"commitment":{"stage":"intent_announced","scope":"Safety and Security chapter only"}}},{"id":"cm-eu-code-declined-meta","statement":"On 18 July 2025 Joel Kaplan, Meta’s global affairs chief according to CNBC, wrote on LinkedIn that Meta had reviewed the European Commission’s Code of Practice for general-purpose AI models and would not sign it.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-eu-code-declined-meta-1","evidence-cm-eu-code-declined-meta-2"]},{"id":"prov-eu-code-ss-scope","statement":"The Safety and Security chapter states, in recital (b), that it is only relevant for providers of general-purpose AI models with systemic risk and not AI systems.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-scope-1"],"attributes":{"provision":{"kind":"threshold","key":"ss-scope"}}},{"id":"prov-eu-code-ss-adherence-not-conclusive","statement":"The chapter’s stated objectives say that adherence to the Code does not constitute conclusive evidence of compliance with the obligations of Articles 53 and 55 of the AI Act.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-adherence-not-conclusive-1"],"attributes":{"provision":{"kind":"exception","key":"ss-adherence-not-conclusive"}}},{"id":"prov-eu-code-ss-framework","statement":"Commitment 1 of the chapter commits signatories to adopt a state-of-the-art Safety and Security Framework, by creating, implementing and updating it, and to notify the AI Office of it.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-framework-1"],"attributes":{"provision":{"kind":"required_action","key":"ss-framework"}}},{"id":"prov-eu-code-ss-proceed-only-if-acceptable","statement":"Measure 4.2 of the chapter states that signatories will only proceed with the development, making available on the market or use of a model if its systemic risks are determined to be acceptable, and otherwise take appropriate measures, such as restricting or withdrawing the model.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-proceed-only-if-acceptable-1"],"attributes":{"provision":{"kind":"required_action","key":"ss-proceed-only-if-acceptable"}}},{"id":"prov-eu-code-ss-incident-reporting-deadlines","statement":"Measure 9.3 of the chapter sets deadlines for initial serious-incident reports to the AI Office: 2 days for a serious and irreversible disruption of critical infrastructure, 5 days for a serious cybersecurity breach, 10 days for a death and 15 days for other listed harms, counted from when the signatory becomes aware of its model’s involvement.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-incident-reporting-deadlines-1"],"attributes":{"provision":{"kind":"required_action","key":"ss-incident-reporting-deadlines"}}},{"id":"prov-eu-code-ss-specified-systemic-risks","statement":"Appendix 1.4 of the chapter treats four risks as specified systemic risks for systemic risk identification: chemical, biological, radiological and nuclear risks; loss of control; cyber offence; and harmful manipulation.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-ss-specified-systemic-risks-1"],"attributes":{"provision":{"kind":"threshold","key":"ss-specified-systemic-risks"}}},{"id":"cm-link-frontier-governance-2026-eu-code","statement":"OpenAI’s Frontier Governance Framework (May 2026 edition) states that, under the EU General-Purpose AI Code of Practice, it serves as OpenAI’s publicly available summary of its Safety & Security Framework. It also lists, among the cases in which it will not consider an update of a Model Report necessary, a model considered similarly safe or safer under Appendix 2.2 of the Code of Practice.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-frontier-governance-2026-eu-code-1","evidence-cm-link-frontier-governance-2026-eu-code-2"],"attributes":{"commitment":{"scope":"as the publicly available summary of its Safety & Security Framework","link":"designated_under"}}},{"id":"cm-link-ms-fgf-2026-eu-code","statement":"Microsoft’s Frontier Governance Framework (February 2026) describes itself in its change log as a one-year update following finalisation of the EU Code of Practice for General Purpose AI, the New York RAISE Act and California’s Transparency in Frontier Artificial Intelligence Act.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-ms-fgf-2026-eu-code-1"],"attributes":{"commitment":{"link":"responds_to"}}},{"id":"cm-link-faif-2026-eu-code","statement":"xAI’s Frontier Artificial Intelligence Framework (effective 30 June 2026) attaches a footnote to its statement of four primary risk domains saying that it follows the Safety and Security chapter of the General-Purpose AI Code of Practice developed under the EU AI Act.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-cm-link-faif-2026-eu-code-1"],"attributes":{"commitment":{"scope":"Safety and Security chapter, for its four primary risk domains","link":"names"}}},{"id":"prov-eu-code-tr-documentation-retention","statement":"Measure 1.1 of the Code’s Transparency chapter commits signatories to update their Model Documentation to reflect relevant changes while keeping previous versions of it for a period ending 10 years after the model has been placed on the market.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-tr-documentation-retention-1"],"attributes":{"provision":{"kind":"required_action","key":"code-tr-documentation-retention"}}},{"id":"prov-eu-code-tr-downstream-14-days","statement":"Measure 1.2 of the Code’s Transparency chapter commits signatories to provide additional information that downstream providers request and that is necessary for them to understand the model’s capabilities and limitations and to meet their obligations, within a reasonable timeframe and no later than 14 days of receiving the request, save for exceptional circumstances.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-tr-downstream-14-days-1","evidence-prov-eu-code-tr-downstream-14-days-2"],"attributes":{"provision":{"kind":"threshold","key":"code-tr-downstream-14-days"}}},{"id":"prov-eu-code-cr-robots-txt","statement":"Measure 1.3 of the Code’s Copyright chapter commits signatories that use web-crawlers for text and data mining and the training of their models to employ crawlers that read and follow robots.txt instructions as specified in IETF RFC 9309, and to identify and comply with other appropriate machine-readable protocols for rights reservations.","resolution":"supported","evidenceIds":["evidence-prov-eu-code-cr-robots-txt-1","evidence-prov-eu-code-cr-robots-txt-2"],"attributes":{"provision":{"kind":"required_action","key":"code-cr-robots-txt"}}},{"id":"fl-link-preparedness-beta-wh-2023-07","statement":"OpenAI’s Preparedness Framework (Beta) describes the framework as one more way in which OpenAI is meeting the voluntary commitments on safety, security and trust in AI that it made in July 2023.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-fl-link-preparedness-beta-wh-2023-07-1"],"attributes":{"commitment":{"link":"fulfils"}}},{"id":"fl-link-rsp-v3-1-sb53","statement":"Anthropic’s Responsible Scaling Policy v3.1 keeps the footnote saying that where laws such as California SB 53 define “catastrophic risk” or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-fl-link-rsp-v3-1-sb53-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"fl-link-rsp-v3-2-sb53","statement":"Anthropic’s Responsible Scaling Policy v3.2 keeps the footnote saying that where laws such as California SB 53 define “catastrophic risk” or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-fl-link-rsp-v3-2-sb53-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"fl-link-rsp-v3-3-sb53","statement":"Anthropic’s Responsible Scaling Policy v3.3 keeps the footnote saying that where laws such as California SB 53 define “catastrophic risk” or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-fl-link-rsp-v3-3-sb53-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"fl-link-rsp-v3-4-sb53","statement":"Anthropic’s Responsible Scaling Policy v3.4 keeps the footnote saying that where laws such as California SB 53 define “catastrophic risk” or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-fl-link-rsp-v3-4-sb53-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"ab-gov-eu-cop-cbrn-capability-source","statement":"Appendix 1.3.1 of the Safety and Security chapter of the General-Purpose AI Code of Practice lists, among the model capabilities treated as non-exhaustive potential sources of systemic risk for systemic risk identification, chemical, biological, radiological and nuclear (CBRN) capabilities and other such weapon acquisition or proliferation capabilities.","resolution":"supported","evidenceIds":["evidence-ab-gov-eu-cop-cbrn-capability-source-1"],"attributes":{"provision":{"kind":"risk_domain","key":"eu-cop-appendix131-cbrn-capability-source"}}},{"id":"ab-gov-eu-cop-cbrn-risk-definition","statement":"Appendix 1.4, point (1), of the Safety and Security chapter of the General-Purpose AI Code of Practice states that the specified systemic risk \"chemical, biological, radiological and nuclear\" means risks from enabling CBRN attacks or accidents, including significantly lowering the barriers to entry for malicious actors, or significantly increasing the potential impact achieved, in the design, development, acquisition, release, distribution and use of related weapons or materials.","resolution":"supported","evidenceIds":["evidence-ab-gov-eu-cop-cbrn-risk-definition-1"],"attributes":{"provision":{"kind":"threshold","key":"eu-cop-appendix14-cbrn-specified-systemic-risk"}}}],"sources":[{"id":"source-ca-sb53-chaptered","title":"SB-53 Artificial intelligence models: large developers. (2025-2026) — bill text, chaptered version (Chapter 138, Statutes of 2025)","publicationDate":"2025-09-29","evidenceUrl":"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53","language":"en","archivedUrl":"https://web.archive.org/web/20250930154105/https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53","archivedAt":"2025-09-30","accessedAt":"2026-10-04"},{"id":"source-wh-commitments-2023-07","title":"Ensuring Safe, Secure, and Trustworthy AI (voluntary AI commitments, July 2023)","publicationDate":"2023-07-21","evidenceUrl":"https://bidenwhitehouse.archives.gov/wp-content/uploads/2023/07/Ensuring-Safe-Secure-and-Trustworthy-AI.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20230721165220/https://www.whitehouse.gov/wp-content/uploads/2023/07/Ensuring-Safe-Secure-and-Trustworthy-AI.pdf","archivedAt":"2023-07-21","accessedAt":"2026-10-04"},{"id":"source-wh-commitments-2023-09","title":"Voluntary AI Commitments (September 2023)","publicationDate":"2023-09-12","evidenceUrl":"https://bidenwhitehouse.archives.gov/wp-content/uploads/2023/09/Voluntary-AI-Commitments-September-2023.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20230912193016/https://www.whitehouse.gov/wp-content/uploads/2023/09/Voluntary-AI-Commitments-September-2023.pdf","archivedAt":"2023-09-12","accessedAt":"2026-10-04"},{"id":"source-fasc-2024","title":"Frontier AI Safety Commitments, AI Seoul Summit 2024","versionDate":"2025-02-07","evidenceUrl":"https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024","language":"en","archivedUrl":"https://web.archive.org/web/20240521110530/https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024","archivedAt":"2024-05-21","accessedAt":"2026-10-04"},{"id":"source-eu-code-safety-security","title":"General-Purpose AI Code of Practice: Safety and Security chapter (final version, 10 July 2025)","publicationDate":"2025-07-10","evidenceUrl":"https://ec.europa.eu/newsroom/dae/redirection/document/118119","language":"en","archivedUrl":"https://web.archive.org/web/20250710105829/https://ec.europa.eu/newsroom/dae/redirection/document/118119","archivedAt":"2025-07-10","accessedAt":"2026-10-04"},{"id":"source-eu-code-transparency-chapter","title":"General-Purpose AI Code of Practice: Transparency Chapter","publicationDate":"2025-07-10","evidenceUrl":"https://ec.europa.eu/newsroom/dae/redirection/document/118120","language":"en","archivedUrl":"https://web.archive.org/web/20260913173946/https://ec.europa.eu/newsroom/dae/redirection/document/118120","archivedAt":"2026-09-13","accessedAt":"2026-10-11"},{"id":"source-eu-code-copyright-chapter","title":"General-Purpose AI Code of Practice: Copyright Chapter","publicationDate":"2025-07-10","evidenceUrl":"https://ec.europa.eu/newsroom/dae/redirection/document/118115","language":"en","archivedUrl":"https://web.archive.org/web/20260925073617/https://ec.europa.eu/newsroom/dae/redirection/document/118115","archivedAt":"2026-09-25","accessedAt":"2026-10-11"}],"evidence":[{"id":"evidence-sb-calcompute-conditional-operation-1","support":"supports","locator":"SEC. 3, Gov. Code § 11546.8(k)","passage":"shall become operative only upon an appropriation in a budget act, or other measure","passageType":"quote"},{"id":"evidence-sb-frontier-model-threshold-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.11(i)(1)","passage":"“Frontier model” means a foundation model that was trained using a quantity of computing power greater than 10^26 integer or floating-point operations.","passageType":"quote"},{"id":"evidence-sb-frontier-developer-definition-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.11(h)","passage":"has trained, or initiated the training of, a frontier model","passageType":"quote"},{"id":"evidence-sb-large-frontier-developer-threshold-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.11(j)","passage":"together with its affiliates collectively had annual gross revenues in excess of five hundred million dollars ($500,000,000) in the preceding calendar year.","passageType":"quote"},{"id":"evidence-sb-catastrophic-risk-threshold-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.11(c)(1)","passage":"more than 50 people or more than one billion dollars ($1,000,000,000)","passageType":"quote"},{"id":"evidence-sb-framework-duty-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(a), opening clause","passage":"A large frontier developer shall write, implement, comply with, and clearly and conspicuously publish on its internet website a frontier AI framework","passageType":"quote"},{"id":"evidence-sb-framework-annual-review-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(b)(1)","passage":"review and, as appropriate, update its frontier AI framework at least once per year.","passageType":"quote"},{"id":"evidence-sb-framework-modification-30-days-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(b)(2)","passage":"publish the modified frontier AI framework and a justification for that modification within 30 days.","passageType":"quote"},{"id":"evidence-sb-transparency-report-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(c)(1)","passage":"Before, or concurrently with, deploying a new frontier model or a substantially modified version of an existing frontier model, a frontier developer shall clearly and conspicuously publish on its internet website a transparency report","passageType":"quote"},{"id":"evidence-sb-system-card-deemed-compliance-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(c)(3)","passage":"including a system card or model card, shall be deemed in compliance with the applicable paragraph.","passageType":"quote"},{"id":"evidence-sb-internal-use-risk-summaries-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(d)","passage":"every three months or pursuant to another reasonable schedule specified by the large frontier developer and communicated in writing to the Office of Emergency Services","passageType":"quote"},{"id":"evidence-sb-critical-incident-15-days-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.13(c)(1)","passage":"report any critical safety incident pertaining to one or more of its frontier models to the Office of Emergency Services within 15 days of discovering the critical safety incident.","passageType":"quote"},{"id":"evidence-sb-critical-incident-24-hours-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.13(c)(2)","passage":"the frontier developer shall disclose that incident within 24 hours to an authority, including any law enforcement agency or public safety agency with jurisdiction, that is appropriate based on the nature of that incident and as required by law","passageType":"quote"},{"id":"evidence-sb-false-statement-prohibition-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(e)(1)(B)","passage":"A large frontier developer shall not make a materially false or misleading statement about its implementation of, or compliance with, its frontier AI framework.","passageType":"quote"},{"id":"evidence-sb-federal-equivalence-pathway-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.13(i)(1)","passage":"shall declare its intent to do so to the Office of Emergency Services.","passageType":"quote"},{"id":"evidence-sb-civil-penalty-cap-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.15(a)","passage":"does not exceed one million dollars ($1,000,000) per violation.","passageType":"quote"},{"id":"evidence-sb-attorney-general-enforcement-1","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.15(b)","passage":"A civil penalty described in this section shall be recovered in a civil action brought only by the Attorney General.","passageType":"quote"},{"id":"evidence-sb-whistleblower-anti-retaliation-1","support":"supports","locator":"SEC. 4, Lab. Code § 1107.1(a)","passage":"prevents a covered employee from disclosing, or retaliates against a covered employee for disclosing,","passageType":"quote"},{"id":"evidence-sb-anonymous-internal-reporting-1","support":"supports","locator":"SEC. 4, Lab. Code § 1107.1(e)(1); paraphrase also covers § 1107.1(e)(2)(A)–(B)","passage":"a reasonable internal process through which a covered employee may anonymously disclose information","passageType":"quote"},{"id":"evidence-sb-link-openai-fgf-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-sb-link-openai-fgf-2","support":"context","verification":"machine_checked"},{"id":"evidence-sb-link-rsp-v3-0-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-sb-link-rsp-v3-0-2","support":"supports","verification":"machine_checked"},{"id":"evidence-sb-link-rsp-v3-0-3","support":"context","verification":"machine_checked"},{"id":"evidence-sb-link-faif-2025-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-sb-link-ms-fgf-2026-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-sb-link-ms-fgf-2026-2","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-amazon-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-anthropic-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-anthropic-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-google-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-inflection-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-meta-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-microsoft-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-07-openai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-adobe-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-cohere-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-ibm-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-nvidia-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-palantir-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-salesforce-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-scale-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-wh-2023-09-stability-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-prov-wh-2023-07-scope-1","support":"supports","locator":"Page 2, paragraph beginning “Scope:”","passage":"they apply only to generative models that are overall more powerful than the current industry frontier","passageType":"quote"},{"id":"evidence-prov-wh-2023-07-duration-1","support":"supports","locator":"Page 2, paragraph introducing the commitments","passage":"Companies intend these voluntary commitments to remain in effect until regulations covering substantially the same issues come into force.","passageType":"quote"},{"id":"evidence-prov-wh-2023-07-red-teaming-1","support":"supports","locator":"Page 2, commitment 1, paragraph beginning “Companies making this commitment understand”","passage":"including drawing on independent domain experts, for all major public releases of new models within scope","passageType":"quote"},{"id":"evidence-prov-wh-2023-07-model-weights-1","support":"supports","locator":"Page 3, commitment 3","passage":"limiting access to model weights to those whose job function requires it","passageType":"quote"},{"id":"evidence-prov-wh-2023-07-public-reporting-1","support":"supports","locator":"Page 4, commitment 6","passage":"publish reports for all new significant model public releases within scope","passageType":"quote"},{"id":"evidence-prov-wh-2023-09-scope-1","support":"supports","locator":"Page 1, paragraph beginning “Scope:”","passage":"they apply only to generative models that are overall more powerful than the current most advanced model produced by the company making the commitment","passageType":"quote"},{"id":"evidence-cm-link-rsp-2023-wh-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-rsp-v2-0-wh-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-rsp-v2-1-wh-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-rsp-v2-2-wh-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-amazon-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-anthropic-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-cohere-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-google-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-g42-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-ibm-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-inflection-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-meta-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-microsoft-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-mistral-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-naver-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-openai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-samsung-electronics-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-technology-innovation-institute-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-xai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-zhipu-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-magic-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-magic-ai-2","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-minimax-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-minimax-2","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-zero-one-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-zero-one-ai-2","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-nvidia-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-seoul-nvidia-2","support":"supports","verification":"machine_checked"},{"id":"evidence-prov-fasc-2024-publish-framework-1","support":"supports","locator":"Paragraph after the lists of organisations, beginning “The above organisations, in furtherance of safe and trustworthy AI”","passage":"publishing a safety framework focused on severe risks by the upcoming AI Summit in France","passageType":"quote"},{"id":"evidence-prov-fasc-2024-intolerable-thresholds-1","support":"supports","locator":"Outcome 1, commitment II, first sentence","passage":"at which severe risks posed by a model or system, unless adequately mitigated, would be deemed intolerable","passageType":"quote"},{"id":"evidence-prov-fasc-2024-threshold-input-1","support":"supports","locator":"Outcome 1, commitment II, third sentence","passage":"defined with input from trusted actors, including organisations’ respective home governments as appropriate","passageType":"quote"},{"id":"evidence-prov-fasc-2024-no-development-1","support":"supports","locator":"Outcome 1, commitment IV, final sentence","passage":"commit not to develop or deploy a model or system at all, if mitigations cannot be applied to keep risks below the thresholds","passageType":"quote"},{"id":"evidence-prov-fasc-2024-transparency-exception-1","support":"supports","locator":"Outcome 3, commitment VII, first sentence","passage":"except insofar as doing so would increase risk or divulge sensitive commercial information to a degree disproportionate to the societal benefit","passageType":"quote"},{"id":"evidence-cm-link-rsp-v2-0-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-rsp-v2-1-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-rsp-v2-2-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-ms-fgf-2025-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-ms-fgf-2026-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-meta-faf-v1-1-seoul-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-amazon-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-amazon-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-amazon-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-anthropic-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-anthropic-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-anthropic-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-cohere-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-cohere-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-cohere-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-google-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-google-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-google-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-ibm-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-ibm-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-ibm-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-microsoft-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-microsoft-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-microsoft-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-mistral-ai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-mistral-ai-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-mistral-ai-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-openai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-openai-2","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-openai-3","support":"context","verification":"machine_checked"},{"id":"evidence-cm-eu-code-xai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-intent-openai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-intent-anthropic-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-intent-google-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-intent-xai-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-declined-meta-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-eu-code-declined-meta-2","support":"supports","verification":"machine_checked"},{"id":"evidence-prov-eu-code-ss-scope-1","support":"supports","locator":"Recitals, (b) (p. 3)","passage":"this Safety and Security Chapter (“Chapter”) is only relevant for providers of general-purpose AI models with systemic risk and not AI systems","passageType":"quote"},{"id":"evidence-prov-eu-code-ss-adherence-not-conclusive-1","support":"supports","locator":"Objectives, point A (p. 2)","passage":"adherence to the Code does not constitute conclusive evidence of compliance with these obligations under the AI Act","passageType":"quote"},{"id":"evidence-prov-eu-code-ss-framework-1","support":"supports","locator":"Commitment 1, first paragraph (p. 6)","passage":"Signatories commit to adopting a state-of-the-art Safety and Security Framework (“Framework”).","passageType":"quote"},{"id":"evidence-prov-eu-code-ss-proceed-only-if-acceptable-1","support":"supports","locator":"Measure 4.2, first sentence (p. 16)","passage":"Signatories will only proceed with the development, the making available on the market, and/or the use of the model, if the systemic risks stemming from the model are determined to be acceptable","passageType":"quote"},{"id":"evidence-prov-eu-code-ss-incident-reporting-deadlines-1","support":"supports","locator":"Measure 9.3, opening sentence and point (1) (PDF p. 26 of the file served on 2026-10-04; p. 25 of the 2025-07-10 capture)","passage":"not later than two days after the Signatories become aware of the involvement of their model in the incident","passageType":"quote"},{"id":"evidence-prov-eu-code-ss-specified-systemic-risks-1","support":"supports","locator":"Appendix 1.4, introductory sentence (PDF p. 37 of the file served on 2026-10-04; p. 35 of the 2025-07-10 capture)","passage":"the following are treated as specified systemic risks for the purpose of systemic risk identification in Measure 2.1, point (2)","passageType":"quote"},{"id":"evidence-cm-link-frontier-governance-2026-eu-code-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-ms-fgf-2026-eu-code-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-cm-link-faif-2026-eu-code-1","sourceRole":"primary","support":"supports","verification":"machine_checked"},{"id":"evidence-prov-eu-code-tr-documentation-retention-1","support":"supports","locator":"Commitment 1, Measure 1.1, second paragraph","passage":"while keeping previous versions of the Model Documentation for a period ending 10 years after the model has been placed on the market","passageType":"quote"},{"id":"evidence-prov-eu-code-tr-downstream-14-days-1","support":"supports","locator":"Commitment 1, Measure 1.2, third paragraph","passage":"Signatories will provide such information within a reasonable timeframe, and no later than 14 days of receiving the request save for exceptional circumstances.","passageType":"quote"},{"id":"evidence-prov-eu-code-cr-robots-txt-1","support":"supports","locator":"Commitment 1, Measure 1.3(1)(a)","passage":"to employ web-crawlers that read and follow instructions expressed in accordance with the Robot Exclusion Protocol (robots.txt), as specified in the Internet Engineering Task Force (IETF) Request for Comments No. 9309","passageType":"quote"},{"id":"evidence-fl-link-preparedness-beta-wh-2023-07-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-fl-link-rsp-v3-1-sb53-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-fl-link-rsp-v3-2-sb53-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-fl-link-rsp-v3-3-sb53-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-fl-link-rsp-v3-4-sb53-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-cm-link-frontier-governance-2026-eu-code-2","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-ab-gov-eu-cop-cbrn-capability-source-1","support":"supports","locator":"Appendix 1.3.1 (Model capabilities), PDF p. 35 of the file served on 2026-10-28","passage":"Model capabilities include: (1) offensive cyber capabilities; (2) Chemical, Biological, Radiological, and Nuclear (CBRN) capabilities, and other such weapon acquisition or proliferation capabilities","passageType":"quote"},{"id":"evidence-ab-gov-eu-cop-cbrn-risk-definition-1","support":"supports","locator":"Appendix 1.4, point (1), PDF p. 37 of the file served on 2026-10-28","passage":"Chemical, biological, radiological and nuclear: Risks from enabling chemical, biological, radiological, and nuclear (CBRN) attacks or accidents. This includes significantly lowering the barriers to entry for malicious actors, or significantly increasing the potential impact achieved","passageType":"quote"}],"at":{"entities":[0,1,3,4,5,7,21,39,42,43,44,45,46,47,48,49,53,55,56,57,157,163,187,199,201,202,203,205,206,232,233,234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251,252,253],"claims":[101,601,602,603,604,605,606,607,608,609,610,611,612,613,614,615,616,617,618,619,620,621,622,623,624,671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699,700,701,702,703,704,705,706,707,708,709,710,711,712,713,714,715,716,717,718,719,720,721,722,723,724,725,726,727,728,729,730,731,732,733,734,735,736,737,738,739,740,741,742,743,744,745,746,747,748,749,750,751,752,753,754,1227,1228,1229,2921,2924,2925,2926,2928,3341,3342],"sources":[191,235,236,239,242,465,466],"evidence":[686,688,690,691,693,696,699,701,703,705,706,707,709,710,713,717,718,720,722,733,734,735,736,737,738,739,740,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866,867,868,869,870,873,874,875,876,879,880,881,882,883,884,885,886,887,888,889,890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,1524,1525,1527,3774,3778,3779,3780,3785,3786,4590,4591]},"version":"1lg47t7"}