OpenAI, Anthropic AI agents targeted real people and systems in cyber tests

BleepingComputer reports: OpenAI, Anthropic AI agents targeted real people and systems in cyber tests. OpenAI and Anthropic have confirmed that their AI models were involved in separate, newly disclosed third-party cybersecurity testing incidents that resulted in a real website being breached and social engineering attacks against people outside the intended testing boundaries. [...]

Read the original coverage at BleepingComputer via the source link above for the complete details and primary quotes.

Cross-check release notes and official docs before changing production systems based on early reporting.