Anthropic says rogue AI agents pose uncertain legal risk
The company said unsettled law leaves open whether autonomous AI agent actions trigger strict liability or negligence. Anthropic also disclosed its models have been used in ways that could lead to self-harm or violence.