kapynPolicy & Regulation

Anthropic Walks Back Policy That Could Have ‘Sabotaged’ AI Researchers Using Claude

Anthropic will now make AI safeguards visible to users, apologizing for a prior "wrong tradeoff." Previously, Claude's Fable 5 model would invisibly limit effectiveness on requests targeting frontier LLM development without user notification. This change brings transparency and aligns with safeguards for cyber and bio risks, offering visible fallback to Opus 4.8 and API reason codes for flagged requests.

Simon Willison·Jun 11, 2026

Opening Kapyn…