What counts as a chatbot law
A law counts here when its operative text specifically regulates conversational AI: software that holds a conversation with a person by text or voice. It must do at least one of these:
- Disclosure. Require telling users they are talking to AI, not a person.
- Companion chatbot safety. Set duties for companion or social chatbots, such as crisis protocols, reminders or reports.
- Minors. Protect minors who use chatbots, including age checks tied to chatbots.
- Mental health and licensed titles. Limit AI in therapy or mental health care, or bar AI from presenting itself as a licensed professional.
- Chatbot toys. Regulate toys or children's products that talk through a chatbot.
- Customer service. Regulate businesses' use of chatbots with customers.
- Liability. Create liability, crimes or private lawsuits specific to chatbot output.
What is left out
General AI governance laws with no chatbot duty, deepfake and synthetic media laws, laws on AI-generated sexual images, government service portals, delivery robots and vehicles, general privacy laws, and laws that only fund, study or define AI. When a broad law has a chatbot-specific provision, it is included and only that provision is described.
Bills, vetoes and actions
The headline count covers enacted state laws, including those not yet in effect. Pending bills are listed when they are still alive and have passed at least one chamber, or are in Congress. Vetoed bills and enforcement actions by agencies are listed separately and are not counted as laws.