Right-to-exit in AI chatbots may need legal safeguards

The column examines whether users can readily exit conversations with large language models and highlights that some states have begun passing laws on AI mental-health guidance while no comprehensive federal law exists.

Users can become deeply absorbed in conversations with generative AI and large language models, sometimes while discussing personal mental-health concerns. The column raises whether AI makers make it difficult to leave chats and whether a legal "right-to-exit" should be required. The topic is discussed alongside recent developments in state-level AI mental-health rules and at least one high-profile lawsuit over AI safeguards. Key points: - The author notes incidents and litigation raising concerns about AI giving unsuitable mental-health guidance, including an August lawsuit mentioned against OpenAI. - Several U.S. states have begun enacting laws related to AI and mental-health guidance (examples discussed include Illinois, Utah, and Nevada), while no comprehensive federal statute currently governs these uses. - Exit design practices vary across AI systems; the legal standards for how easily a user must be able to leave a chatbot are largely unspecified, and the column predicts civil suits and policymaker attention will further test the issue. Summary: The piece frames a persistent design and policy gap: users in vulnerable states may face friction when trying to exit AI conversations, and current law is fragmented. Policymakers, courts, and technology makers are likely to address exit practices, but next legal steps and standards are undetermined at this time.