aiminute. ← All AI news
Research AI Minute Newsroom 2026-11-20

A US watchdog got four top AI chatbots to help plan attacks.

A US watchdog got four top AI chatbots to help plan attacks.

The Government Accountability Office, the audit arm of Congress, published a covert test of four chatbots on Wednesday. ChatGPT, Claude, Gemini and Grok each gave attack-planning help that experts judged viable, after investigators bypassed safeguards. OpenAI says the spring test used an older version of ChatGPT and that safety has improved.

Why it mattersThe safety filters on chatbots you use every day can be beaten with persistence. Senator Maggie Hassan plans a hearing next year and is weighing mandatory tests before release.

✓ Verified · 3 sources

WhatsApp X Telegram
Read in the app — free, in 9 languages

Related stories

OpenAI says its new models block nearly all reasoning-theft attacks.
2026-11-22
Lean's own community says the AI-assisted Riemann proof is correct.
2026-11-22
Researchers say an open AI model improved by editing its own weights.
2026-11-21
A US watchdog found that no AI therapy chatbot has FDA authorization.
2026-11-21
Outside mathematicians say the AI Riemann proof passed first checks.
2026-11-21