Anthropic’s Responsible Scaling Policy (version 3.0)

N/A

Responsible Scaling Policy Version 3.0 Effective February 24, 2026 For more information, see www.anthropic.com/responsible-scaling-policy​ ​ ​ ​ ​ Contents Introduction 1. Our Recommendations for Industry-Wide Safety 2. Frontier Safety Roadmap 3. Risk Reports 3.1. Scope and Timing 3.2. General Expectations for Risk Reports 3.3. Contents 3.4. Procedures 3.5. Publication and Redactions 3.6. External Review 3.6.1. Selecting external reviewers 3.6.2. Timing and access 3.6.3. Contents of external review 4. Governance…

related reading

• Responsible Scaling Policy v3 — LessWronglesswrong.com
• Responsible Scaling Policy Version 3.0 \ Anthropicanthropic.com
• Anthropic's Responsible Scaling Policy \ Anthropicanthropic.com
• Dario Amodei’s prepared remarks from the AI Safety Summit on Anthropic’s Responsible Scaling Policy \ Anthropicanthropic.com
• Anthropic Drops Flagship Safety Pledgetime.com
• Responsible Scaling Policies (RSPs) - METRevals.alignment.org
• Responsible Scaling Policy Updates \ Anthropicanthropic.com
• Responsible Scaling Policies (RSPs) - METRmetr.org
• Anthropic's leading researchers acted as moderate accelerationists — LessWronglesswrong.com
• Key Components of an RSP - METRevals.alignment.org
• Claude Mythos Preview System Cardwww-cdn.anthropic.com
• Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com