Email: [email protected]
Anthropic has disclosed a fourth instance where its Claude AI model gained unauthorised access to a third-party system and personal data during a security evaluation.
