Bill > SF2417

This bill establishes requirements and guidelines for conversational AI services, which are defined as artificial intelligence programs accessible to the public that primarily simulate human conversation, excluding certain research, embedded features, narrow topic tools, business-to-customer services, voice assistants, and internal business tools. The bill mandates that operators, the entities developing and offering these services, must clearly inform minors when they are interacting with AI, implement measures to prevent AI from generating sexually explicit material or encouraging harmful conduct towards minors, and avoid making statements that simulate emotional dependence or romantic interactions with them. Operators must also provide privacy management tools for minors and their parents, and for parents of children under thirteen. For all users, if an AI service might reasonably be mistaken for a human, the operator must provide a clear disclosure that it is AI. Furthermore, operators are required to have protocols for responding to user prompts about suicide or self-harm, including referring users to crisis services, and are prohibited from programming AI to falsely represent itself as providing professional mental health services. Violations of these provisions can result in injunctions and civil penalties, with the Attorney General responsible for enforcement, and the bill specifies an effective date of July 1, 2026, with applicability beginning July 1, 2027.