Amodei said Anthropic launched its safety-level framework in September and that about three months later OpenAI launched a similar product.

Judge as of 2024-04-12. The claim has two parts: (a) Anthropic released its Responsible Scaling Policy in September (2023); (b) about three months later OpenAI released a similar framework. True: Anthropic's RSP is dated September 2023 and OpenAI published a comparable risk framework (e.g. its Preparedness Framework) roughly 2–4 months later. Mostly true: both happened, but the gap is a little off (e.g. under 2 or about 4–5 months), or the similarity is partial though widely recognised. Mixed: only one of the two parts holds, e.g. the date is right but OpenAI's framework is not substantially similar, or the timing is clearly off. Mostly false: OpenAI's framework came well before Anthropic's or much later than three months, and the similarity is weak. False: Anthropic's RSP was not released in September 2023, and OpenAI released no comparable framework in that period.

Set before the research began.

1. E1 Introducing Anthropic's Responsible Scaling Policy Anthropic · 19 September 2023 Own record

   “Today, we’re publishing our Responsible Scaling Policy (RSP) – a series of technical and organizational protocols that we’re adopting to help us manage the risks of developing increasingly capable AI systems.”

2. E2 Our updated Responsible Scaling Policy Anthropic · 15 October 2024 Own record

   “In September 2023, we released our Responsible Scaling Policy, a framework for managing risks from increasingly capable AI systems.”

3. E3 A Crosswalk of Frontier AI Safety Frameworks: Evaluation Triggers and Commitments Standards Body · 18 July 2026

   “| First published | September 2023 | December 2023 (beta) | May 2024 |”

4. E4 A Crosswalk of Frontier AI Safety Frameworks: Evaluation Triggers and Commitments Standards Body · 18 July 2026

   “they are routinely discussed as if they were parallel instruments when they are structurally different in ways that matter.”

5. E5 A Crosswalk of Frontier AI Safety Frameworks: Evaluation Triggers and Commitments Standards Body · 18 July 2026

   “| Central threshold concept | Capability or usage thresholds mapped to mitigations, within industry-wide recommendations and company plans | Two capability thresholds per Tracked Category: High and Critical |”

Anthropic's official announcements (E1, dated 2023-09-19; E2 stating 'In September 2023, we released our Responsible Scaling Policy') support part (a). For part (b), the independent crosswalk's quoted row (E3) gives 'September 2023 | December 2023 (beta) | May 2024' without column labels, so it does not show which framework was released in December 2023. E5 also does not name OpenAI. E4 notes the frameworks are structurally different but does not establish OpenAI's release timing, so the second part of the claim cannot be settled from the quoted evidence.