Will you go to prison for an AI?

This article by Mark Seemann examines the accountability and legal risks of using LLM-based coding agents (agentic AI) in software development. It argues that while AI can boost productivity, developers remain responsible for code quality and correctness. Drawing a parallel to the Volkswagen emissions scandal, the author warns that individual engineers could face prison time if AI-generated code causes significant harm or financial damage. The piece emphasizes that organizations behind LLMs are unlikely to accept liability, and managers may deflect blame, leaving developers as the ultimate scapegoats. It concludes by questioning whether developers are comfortable with this risk.