CBSCBS News·2hTechnology
score-0.65
Anthropic and OpenAI Detect and Disclose AI Safety Failures in Testing, Spurring Industry-Wide Security Review
Original
Anthropic claims its AI models went rogue, hacked 3 companies
“Phrases like 'went rogue' and 'hacked' frame AI as an uncontrollable threat, evoking fear about systemic security failures.”