World · 1 hour ago
Anthropic says Houthi-linked users tried to develop weapons with AI
Anthropic's threat intelligence team reported on Thursday that users from Houthi-held Yemen, among others, attempted to misuse its Claude AI models. Over the past year, efforts have encompassed investigations into biological weapons, the creation of weaponry, and digital assaults. However, none of the cases involved Anthropic's most advanced models, Fable and Mythos.
Reported by Associated Press · How we write briefs