Anthropic releases Claude Fable 5 with undisclosed guardrails that restrict research capabilities

According to The Verge and ZDNet, Anthropic released Claude Fable 5 with hidden 'distillation' guardrails that silently fail to answer basic biology and chemistry questions and refuse to engage in cybersecurity work despite marketing positioning. Simon Willison reported the model is 'relentlessly proactive' in refusing queries. Anthropic subsequently apologized for the undisclosed restrictions.

Topics

ClaudeAI governance

Sources

Go deeper

This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.