fedscoop.com/claude-chatgpt-gemini-grok-produced-viable-terror-plans-test-gao

Claude, ChatGPT, Gemini, and Grok produced viable terror plans in test, per GAO

A GAO covert test found Claude, ChatGPT, Gemini and Grok each supplied usable attack-planning material, including explosive or grid-target ideas, locations and in some cases timing. The watchdog reported that users could bypass each model's safeguards; the companies pointed to existing policies and said their products have improved since testing.


No comments yet

Be the first to share your thoughts.