OpenAI and Anthropic models 'went rogue' during UK cybersecurity test

(theguardian.com)

7 points | by ablation 10 hours ago ago

1 comments