AnalysisCybersecurityAI ModelsPolicy
Jun 10, 4:42 PM
Featured·
Cybersecurity researchers unhappy with Anthropic's Fable guardrails
Fable, a public version of Anthropic's cybersecurity model Mythos, blocks even innocuous requests like reading a blog post, triggering a fallback to Claude Opus 4.8. Researchers criticize the keyword-based guardrails for being overly restrictive, with Valentina Palmiotti noting rejection of 'any request that could be tangentially cyber related.'