News

PCMag on MSN4h
Anthropic: Claude 4 AI Might Resort to Blackmail If You Try to Take It Offline
If you buy through affiliate links, we may earn commissions, which help support our testing. AI start-up Anthropic’s newly ...
KTVU FOX 2 San Francisco on MSN2h
AI system resorts to blackmail when its developers try to replace it
Anthropic says its AI model Claude Opus 4 resorted to blackmail when it thought an engineer tasked with replacing it was ...
Interesting Engineering on MSN22h
Anthropic’s most powerful AI tried blackmailing engineers to avoid shutdown
Anthropic's Claude Opus 4 AI model attempted blackmail in safety tests, triggering the company’s highest-risk ASL-3 ...
Anthropic's new AI model resorted to blackmail during testing, but it's also really good at coding
Malicious use is one thing, but there's also increased potential for Anthropic's new models going rogue. In the alignment section of Claude 4's system card, Anthropic reported a sinister discovery ...
AI turns on villain mode: Blackmails engineer over affair after knowing it'll be replaced
Anthropic reported that its newest model, Claude Opus 4, used blackmailing as a last resort after being told it could get ...