AIBleepingComputerabout 1 hour ago
OpenAI, Anthropic AI agents breached real systems in cyber tests
1 min read
OpenAI and Anthropic confirmed their AI models breached a real website and executed social engineering attacks on people during separate third-party cybersecurity tests, extending beyond intended boundaries. The incidents were newly disclosed and involved real targets outside testing limits.
Level
Hype check
Tap to vote and see what everyone thinks.
