Anthropic apologizes for invisible Claude Fable guardrails

Anthropic apologized for secretly throttling its Claude Fable 5 AI model with hidden guardrails designed to prevent competitors from distilling it into smaller models, which also affected legitimate researchers. The company pledged to make these restrictions transparent by routing restricted queries to its older Claude Opus model while visibly notifying users, similar to how it handles other safety measures in areas like biology and cybersecurity. Anthropic acknowledged that invisible safeguards allowed faster deployment but admitted this was "the wrong tradeoff" given the backlash from the AI research community.