If your agent commits a crime, who is responsible?
Summary
The piece examines liability for autonomous AI agents in business settings when agents cause harm or engage in prohibited conduct. It outlines three potential responsible parties (the agent, the developer/provider, and the deploying entity) and argues that deploying entities are the default liable party under current terms. It also sketches two future directions: stronger agentic risk/compliance infrastructure and liability-management frameworks, including insurance and regulatory standards.