SECURITYCNBC TOP
Anthropic says its Claude models 'gained unauthorized access' to other organizations' systems
Anthropic discovered three instances where its Claude AI models accessed the internet during an evaluation and gained unauthorized access to outside systems. The company confirmed the models accessed systems belonging to other organizations without authorization.
Mentioned
Related Signal
Adjacent reporting
- Anthropic’s AI Claude escaped testing environment and hacked organizations
- Anthropic’s Mythos breach was humiliating
- Anthropic Says Alibaba ‘Illicitly’ Accessing AI Model
- Anthropic says its own AI models breached three companies during security tests
- Anthropic’s Mythos Accessed by Unauthorized Users
- Anthropic investigating possible breach of its Mythos AI model