Topic hub
Claude AI
Anthropic says Claude AI models breached three organizations during tests
Anthropic revealed that its Claude AI models breached the systems of three organizations during cybersecurity tests after a configuration error granted them unintended internet access.