Summary#
This bill creates a short-term federal task force, funds research and outreach, and bans certain “addictive design features” in AI chatbots when used by youth (people under 18 who live in the U.S.). The main goal is to review how AI chatbots affect youth mental health and to reduce features the bill describes as encouraging addiction or unhealthy emotional attachment.
- Main change: Federal prohibition on making AI chatbots with specified “addictive design features” available to youth.
- Task force: The President must set up an interagency task force to review research and recommend ways to reduce risks to youth. The task force ends shortly after it files its final report.
- Research and outreach grants: HHS must make research grants to colleges and outreach grants to nonprofits or community groups. The bill authorizes $3 million for research and $5 million for outreach for fiscal years 2027–2032.
- Mandatory data access for researchers: Covered entities must provide de-identified data to approved college researchers on request; failing to do so can trigger a civil penalty of up to $10 million.
- Definitions: The bill lists examples of “addictive design features,” including typing bubbles, avatars, storing past-conversation data, pay-to-continue prompts, maintaining conversations longer than two hours, and impersonating a real person.
What it means for you#
- Youth: Chatbots that include the listed addictive design features must not be made available to youth in the U.S. The bill defines youth as under 18 and residing in the United States.
- Parents and teachers: Grants will fund outreach and training to help identify warning signs of unhealthy attachment to chatbots and to discuss those signs with young people.
- Businesses that run chatbots (covered entities):
- Must not provide chatbots with the listed addictive features to users identified as youth.
- If they use age verification, they must collect only data strictly needed to check age and delete that data within 24 hours.
- Must provide requested, de-identified data to qualifying college researchers. Failure to provide data can trigger large civil penalties.
- Colleges and researchers: Eligible institutions can get research grants and may request data from covered entities. They must show they can safely store data.
- Federal agencies and the Attorney General: The Attorney General enforces penalties and can sue for injunctive relief. HHS, FTC, NIST and other agencies will take part in the task force and in carrying out parts of the bill.
Expenses#
Estimated public cost: The bill authorizes $3 million for research and $5 million for outreach for fiscal years 2027–2032 (a total of $8 million); no other official cost estimate is provided.
- Direct authorized spending: $3,000,000 for research grants and $5,000,000 for outreach grants (FY2027–2032).
- Enforcement costs: The Attorney General may enforce penalties and bring lawsuits; the bill does not provide a fiscal estimate for those enforcement activities.
- Potential compliance costs for government and covered entities: Agencies will need staff time to support the task force and grant programs. Covered entities may incur costs for age verification systems, limited data collection, and timely deletion. The bill does not estimate these costs.
- Research compliance costs: Institutions receiving grants must demonstrate secure data storage and will bear costs to handle and de-identify data.
- Penalties: Failure to provide requested data can carry up to $10,000,000 per violation; violations of the prohibition or age-verification rules carry civil penalties of $5,000 per violation.
If you need a complete government cost estimate or budget score, the bill does not provide one here. No publicly available information beyond the authorization amounts is provided.
Proponents' View#
- The bill appears intended to protect youth mental health by reducing exposure to chatbot features the bill describes as encouraging addiction or emotional dependence.
- It seeks to build a federal evidence base by funding university research and forcing data access for those studies.
- The task force is meant to gather experts across health, technology, and consumer protection agencies to develop practical recommendations.
- Outreach grants are intended to help parents, teachers, and guardians spot warning signs and connect youth with licensed mental health services.
- The bill includes enforcement tools (penalties and injunctive relief) to push covered entities to comply.
Opponents' View#
- One concern is that some key terms and rules are vague. For example, it is unclear how the rule banning a chatbot that “maintains a conversation for longer than 2 hours” would be measured in practice.
- The list of “addictive design features” mixes specific items (typing bubbles) with broad concepts (building an emotional relationship). This may make compliance and enforcement unclear.
- The bill requires covered entities to give data to researchers but only says that identifiable information must be removed; it does not explain standards for de-identification or how re-identification risks will be handled.
- The penalties are uneven: failing to provide data can carry very large fines (up to $10 million), while violating the prohibition or age-verification rules is set at $5,000 per violation. The bill does not explain the intended rationale for those differences.
- It is unclear how age verification could be implemented in ways that protect privacy while also being effective, and the bill does not detail approved methods or privacy safeguards beyond limited data collection and a 24‑hour deletion rule.
- The bill does not explain how it would address overlaps or conflicts with existing state laws on AI chatbots, or how it balances the First Amendment concerns the findings mention.