Responsible Scaling Policy Version 3.0
Summary
Anthropic announces Version 3.0 of its Responsible Scaling Policy, detailing updated safeguards, a Frontier Safety Roadmap, and periodic Risk Reports to guide AI risk mitigation.
Key quotes
We’re releasing the third version of our Responsible Scaling Policy (RSP), the voluntary framework we use to mitigate catastrophic risks from AI systems.
The new RSP has three key elements.
Risk Reports will be published online (with some redactions) every 3-6 months.
We remain convinced that effective government engagement on AI safety is both necessary and achievable, and we aim to continue advancing a conversation grounded in evidence, national security interests, economic competitiveness, and public trust.
The post explains why Anthropic is revising its policy after two years, evaluates successes and shortcomings of earlier versions, and introduces three new components: separated company‑versus‑industry mitigations, a Frontier Safety Roadmap, and regular Risk Reports with external review.