Is ChatGPT a Product? OpenAI Faces a New Kind of Product Liability Lawsuit

OpenAI is facing a growing wave of lawsuits that could force courts to answer a question traditional product-liability law was never designed for: can an artificial intelligence chatbot be legally treated as a product?

Around two dozen lawsuits have been consolidated in federal court in San Francisco. Plaintiffs allege that ChatGPT’s design and responses contributed to serious harms, including mental-health crises, self-harm and violence. Some of the claims are framed under product-liability theories, alleging that the system was defectively designed or lacked adequate safeguards.

OpenAI disputes the allegations and argues that ChatGPT is an interactive service rather than a conventional product. That distinction could be critical. Product-liability law has traditionally focused on tangible goods with alleged defects in their design, manufacturing or warnings. Applying those principles to software that generates different responses for individual users presents a very different legal challenge.

The litigation also raises questions about whether AI-generated responses should receive protection as speech and whether existing liability rules are capable of addressing technology that can influence users in real time.

No court has yet established that ChatGPT is legally a defective product. But as these cases move forward, their outcomes could help define who bears responsibility when an AI system causes—or allegedly contributes to—real-world harm.