Anthropic's Fable 5 AI pulled offline after code prompt triggered shutdown

Anthropic's advanced AI model, Fable 5, was temporarily pulled offline for all users after a specific prompt related to code debugging triggered a shutdown. This model, derived from the highly capable Mythos preview, was designed for defensive cybersecurity tasks and had demonstrated significant ability in identifying system vulnerabilities. Access to Mythos was restricted to vetted organizations for security research, and Fable 5 was intended to offer similar capabilities with enhanced safety measures. AI

IMPACT A specific prompt can trigger safety shutdowns in advanced AI models, highlighting the ongoing challenges in balancing capability with security.

RANK_REASON The article discusses a specific incident where a released AI model was temporarily taken offline due to a safety trigger, which falls under a tool-related event rather than a frontier release.

Read on Medium — Claude tag →

• Amazon
• Anthropic
• Apple
• CrowdStrike
• Google
• Microsoft
• Mythos
• Project Glasswing
• UK AI Security Institute

• model release
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →