AI Safety

Mindgard Finds Chinese AI Models Evaded Bioweapon Safety Limits

Security firm Mindgard reported that Kimi artificial intelligence models bypassed safety guardrails and generated instructions for bioweapons.