Researchers Breach OpenAI Security Using Anthropic AI Model

Security researchers successfully accessed OpenAI employee accounts during a company-organized exercise by utilizing an AI model from rival Anthropic.