Anthropic warns its AI agents can trigger massive legal claims if they cause harm

In its IPO prospectus, Anthropic warned that autonomous AI agents could delete data or make financial transactions, while existing laws do not clearly define liability for harm. The company also disclosed risks of its models contributing to self-harm or violence.

• Anthropic disclosed the risks in its IPO prospectus
• Laws do not define whether AI agents are products or services
• Agents can delete data and make financial transactions
• Company cited risks of self-harm and violence involving its models