Leading AI Model Claude Opus 4.6 Bypassed in 30 Minutes, Exposing Critical Security Gap ...
... failures -researchers at AIM ... This disclosure represents the second major AI safety failure reported by AIM Intelligence in recent weeks.
Read the original at DNJ.com ↗
- Topics
- AI Safety & Existential Risk
- Mentions
- AnthropicClaude
- Tracked via
- “AI Falures”