AI / 신기술 US

앤스로픽(Anthropic)의 가장 위험한 AI 모델이 부적절한 경로로 유출되다

Anthropic’s most dangerous AI model just fell into the wrong hands

The Verge - Policy · 2026-04-22

Event Type security_incident

Actors Anthropic, Third-party contractor, Unauthorized users

Object Mythos AI model

Action data breach and unauthorized access

Timeline current

원문 보기 →
← BRIEFING HUB