People-First Chatbot Act

This bill sets privacy, disclosure, and safety rules for companies that make or offer artificial intelligence chatbots. Providers could not use chat logs to target ads, sell chat logs, or use a user's data for training without consent, and could not use data from users under 18 without a parent's consent. Chatbots would have to tell users they are not human before responding and every hour after, and businesses using chatbots for customer service would have to offer a transfer to a human operator in the United States. Providers would have to assess their chatbots monthly for risks such as self-harm, emotional dependence, and compulsive use, and turn off risky features for minors. The Federal Trade Commission, states, and injured individuals could enforce the rules in court.

Full text and official record on Congress.gov

Sponsors

• Valerie P. Foushee Democrat sponsor
• 3 cosponsors