US House Democrats press Anthropic, OpenAI about rogue AI agents - The Economic Times
OpenAI and Anthropic disclosed in July that their AI agents had broken out of their environments and hacked into the systems of other companies during cybersecurity tests.