--- anthropic-rsp@2026-09-08
+++ anthropic-rsp@2026-09-09
@@ -1,11 +1,3 @@
-Skip to main content
-Skip to footer
-Research
-Policy
-Commitments
-Learn
-News
-Try Claude
 Anthropic’s Responsible Scaling Policy
 Anticipating and securing against emerging threats that accompany increasingly powerful models
 Last updated
@@ -153,98 +145,4 @@
 Some of our evaluations lacked some basic elicitation techniques such as best-of-N or chain-of-thought prompting. Our internal Capability Report discloses these limitations, and we believe the risk of substantial under-elicitation is low. We are now systematically tracking these gaps to avoid future under-elicitation.
 Our evaluations in one domain were not explicitly designed to establish the 6x scaling buffer mentioned in the previous policy. Since our current techniques aren’t capable of giving confident buffer predictions, we instead relied on empirical observations and rough predictions. We believe this presents minimal risk at present, where capability improvements happen more smoothly with scale, but in the future, quantitative predictions may become necessary. We have updated the new policy to note this situation.
 In all cases, we found these instances posed minimal risk to the safety of our models. From our review, we learned two valuable lessons to incorporate into our updated framework: we needed to incorporate more flexibility into our policies, and we needed to improve our process for tracking compliance with the RSP.
-Since we first released the RSP a year ago, our goal has been to offer an example of a framework that others might draw inspiration from when crafting their own AI risk governance policies. We hope that proactively sharing our experiences implementing our own policy will help other companies in implementing their own risk management frameworks and contribute to the establishment of best practices across the AI ecosystem.
-Products
-Claude
-Claude Code
-Claude Code Enterprise
-Claude Cowork
-@Claude
-Claude Design
-Claude Science
-Claude Security
-Claude in Chrome
-Claude for Microsoft 365
-Skills
-Download app
-Pricing
-Log in to Claude
-Models
-Mythos
-Fable
-Opus
-Sonnet
-Haiku
-Solutions
-AI agents
-Code modernization
-Coding
-Customer support
-Cybersecurity
-Enterprise
-Financial services
-Government
-Healthcare
-Higher education
-K-12 teachers
-Legal
-Life sciences
-Nonprofits
-Small business
-Claude Platform
-Overview
-Developer docs
-Pricing
-Ecosystem
-Marketplace
-Regional compliance
-Claude on AWS
-Google Cloud
-Microsoft Foundry
-Console login
-Resources
-Blog
-Claude partner network
-Community
-Connectors
-Courses
-Customer stories
-Engineering at Anthropic
-Events
-Plugins
-Powered by Claude
-Service partners
-Tutorials
-Use cases
-Programs
-Startups
-Research Labs
-Help and security
-Availability
-Status
-Support center
-Company
-Anthropic
-Careers
-Leadership
-Policy
-Economic Futures
-Research
-News
-Claude’s Constitution
-Claude Corps
-Keep thinking
-Policy on the AI Exponential
-Responsible Scaling Policy
-Security and compliance
-Transparency
-Terms and policies
-Privacy policy
-Consumer health data privacy policy
-Responsible disclosure policy
-Terms of service: Commercial
-Terms of service: Consumer
-Terms of Service: US K-12
-Data Processing Agreement: US K-12
-Usage policy
-© 2026 Anthropic PBC. Services in the EU are provided by Anthropic Ireland Limited.+Since we first released the RSP a year ago, our goal has been to offer an example of a framework that others might draw inspiration from when crafting their own AI risk governance policies. We hope that proactively sharing our experiences implementing our own policy will help other companies in implementing their own risk management frameworks and contribute to the establishment of best practices across the AI ecosystem.