External Publication
Visit Post

Anthropic reportedly 'lost control' of its most dangerous AI model — and that should worry everyone

Tom's Guide | Tech Product Reviews, Top Picks and How To [Unoff… April 22, 2026
Source
Anthropic is facing questions after reports claimed unauthorized users accessed Claude Mythos, its highly restricted cybersecurity AI model. Here’s why it matters.

Discussion in the ATmosphere

Loading comments...