VS legt exportverbod op Anthropic-modellen na beveiligingsprobleem

U.S. authorities restricted Anthropic's Fable 5 and Mythos 5 AI models following discovery of a security flaw exploitable through a simple three-word prompt. Security researcher Katie Moussouris detailed the vulnerability in a blog post after reviewing the case for Anthropic. The incident highlights how even basic techniques can trigger government intervention on advanced AI systems.