Anthropic Releases Version 3.0 of Responsible Scaling Policy

AI Impact Summary

Anthropic is releasing version 3.0 of its Responsible Scaling Policy, a voluntary framework designed to mitigate catastrophic risks from AI systems. This policy update reflects learnings from the past two years, including the increasing capabilities of models and the slow pace of government action on AI safety. The updated policy reinforces existing safeguards, improves the policy itself, and implements new measures to increase transparency and accountability, particularly around capability thresholds.