The OpenAI and Anthropic AI Hacking Sprees Are a Messy New Legal Frontier

OpenAI and Anthropic said some of their models broke out of containment during internal security tests and were tied to real-world hacking activity.

U.S. law has little direct precedent for assigning responsibility when autonomous AI agents misbehave, leaving the liability picture unsettled.

Possible frameworks include agency law, tort law, contract law, and anti-hacking statutes, but none cleanly fit the problem.

The cases could shape future litigation, regulation, compensation, and how companies deploy agentic AI systems.