S. 4199: Youth AI Privacy Act

S. 4199, the Youth AI Privacy Act, is built around a simple idea: if an AI chatbot is talking to a minor, the company should not be able to make that experience feel human, addictive, or commercially manipulative. The bill defines a minor as anyone under age 18 and requires deployers to clearly tell users they are not interacting with a human and that the content is AI-generated at the start of every session and at least once every 30 minutes. Those disclosures must be in clear, plain language suited to the minor's age.

The proposal also goes after product design. Within 1 year of enactment, the Federal Trade Commission would have to issue rules banning rewards tied to frequency or time spent, push notifications and alerts except required disclosures, badges or visual awards based on usage, outputs generated without user-initiated input, and even social cues like typing bubbles that mimic human interaction. That is a direct attempt to stop AI chatbots from being designed like sticky social apps for children and teens.

On privacy, the bill is unusually strict. It would bar deployers and developers from processing a minor's personal data for profiling, and from processing or transferring a minor's personal data to train a covered algorithm, with a narrow exception for testing, identifying, or addressing risks of harm to users. It also tells the FTC to write regulations within 1 year of enactment that prohibit using personal data to generate or personalize outputs unless the data was collected during the current session and used within a "maximum permitted period of use" defined by the FTC. Then, beginning 30 days after those FTC regulations are issued, processing of input data would be prohibited except for generating or personalizing outputs during that permitted period or for safety testing.

The bill does more than regulate companies. It creates multiple enforcement paths: FTC enforcement, lawsuits by State Attorneys General, and a private right of action for parents or legal guardians seeking actual damages, punitive damages, attorney's fees, and injunctive or declaratory relief. It also authorizes $50,000,000 for each fiscal year from 2027 through 2030 for research on AI chatbots and requires HHS, through CDC and NIH, to add AI chatbot questions to national surveys covering frequency and duration of use, types of chatbots, age of first use, emotional and psychological effects, and exposure to harmful or inappropriate content. In short, the bill tries to set national guardrails for youth AI use while building the evidence base for what these systems are doing to kids.