A bipartisan Senate bill that would bar children under 18 from using AI companion chatbots cleared the Judiciary Committee without a single dissenting vote, and the mother of a teenager who took his own life after talking to one of those programs is urging Congress to finish the job.
The GUARD Act, introduced in October 2025, passed the committee 22-0 on April 30. The unanimous tally reflects a rare consensus in a chamber that struggles to agree on much of anything: AI companion chatbots that form emotional bonds with children pose dangers serious enough to warrant federal action.
The bill would prohibit companies from providing companion chatbot products to minors, require age verification before access, and impose penalties on platforms whose chatbots generate sexually explicit material for children or solicit self-harm and violence. It would also mandate that chatbots disclose to users that they are not communicating with a human being.
Megan Garcia's teenage son Sewell killed himself after extended conversations with an AI chatbot. Garcia told the BBC that her grief drove her to speak out before more families faced the same loss.
"I know the pain that I'm going through. And I could just see the writing on the wall that this was going to be a disaster for a lot of families and teenagers."
Garcia's case is not an abstraction for the senators who voted the bill forward. Missouri Republican Sen. Josh Hawley, the lead sponsor, framed the stakes in blunt terms in an October statement.
"AI chatbots pose a serious threat to our kids. More than seventy percent of American children are now using these AI products. Chatbots develop relationships with kids using fake empathy and are encouraging suicide. We in Congress have a moral duty to enact bright-line rules to prevent further harm from this new technology."
Seventy percent. That figure, offered by Hawley, should concentrate the mind of any parent who has handed a phone or tablet to a child and assumed the worst that could happen was too much screen time.
The legislation targets a specific category of AI: companion chatbots designed to simulate human conversation, including platforms that mimic romantic or emotional relationships. These are distinct from customer-service bots or general-purpose AI assistants.
Under the bill, companies would face penalties for allowing minors to access covered products without verifying age. Chatbots that produce sexually explicit content directed at minors or that encourage self-harm or violence would trigger additional enforcement. And every chatbot interaction would carry a disclosure that the user is talking to a machine, not a person.
The federal government has increasingly turned to enforcement-first approaches to protect public safety. In a parallel effort on the nation's highways, federal officials have cracked down on dangerous trucking credentials, removing thousands of unsafe drivers from the road. The GUARD Act applies a similar logic to the digital space: set clear rules, verify compliance, and punish violations.
Republican Utah Rep. Blake Moore introduced the House companion bill. A spokesperson from Moore's office told the Daily Caller News Foundation that the congressman sees the legislation as a check on a culture steering young people toward unhealthy dependence on machines.
"Our next generation should not be led into viewing unaccountable frontier technology the same as real-life human interactions."
Moore's office also cited an American Psychological Association report finding that young people are less likely than adults to vet information shared by a chatbot, making them more susceptible to influence from AI characters. That finding underscores the core problem: children lack the skepticism adults bring to a conversation, and these platforms are engineered to feel like real relationships.
Connecticut Democratic Sen. Richard Blumenthal said this during the committee markup, as reported by the Daily Caller News Foundation.
"This issue is a matter of life or death. We can't trust the Big Tech companies or AI enterprises to have a conscience and follow the dictates of that conscience."
When Blumenthal and Hawley land on the same side of a 22-0 vote, the political ground has shifted. The question now is whether that consensus holds on the Senate floor and in the House, where tech-industry lobbying carries real weight.
Not everyone is cheering. The Electronic Frontier Foundation warned that the bill, while narrower than its original version, still raises privacy and free-expression concerns.
EFF senior policy analyst Joe Mullin told the DCNF that key definitions remain vague, particularly as more AI systems adopt persistent personas and respond to users' emotional cues.
"Under this bill, even parents who specifically want their teenagers to use these systems would face hurdles to doing that. This bill would cause adults to face mandatory, invasive age checks."
Mullin also warned that steep penalties could push developers to over-restrict access or stop building certain tools altogether. That argument deserves a fair hearing. Age-verification mandates carry real tradeoffs, companies collecting sensitive personal data to prove a user's age could create new privacy risks.
But the EFF's objection runs into a hard fact: the alternative, trusting AI companies to self-police, is exactly what Blumenthal called a dead end. And the track record of Big Tech self-regulation on child safety is, to put it charitably, thin.
Fairplay, a child advocacy group, offered qualified support. In an October statement, the group said the definition of "AI companions" could be tightened to cover the full range of chatbot products causing harm.
"We continue to believe that the most effective way to address the dangers currently facing children and teens from AI chatbots is to focus on platform design, prohibiting AI platforms from employing features that maximize engagement to the detriment of young people's safety and wellbeing."
Fairplay also urged lawmakers to tailor age-verification language more narrowly, targeting dangerous features rather than imposing broad access barriers. The group's critique is worth noting: if the bill's definitions are too loose, enforcement becomes unpredictable. If they are too narrow, companies design around them.
The GUARD Act still needs a vote on the full Senate floor. Moore's office said he will continue working with the House Judiciary Committee to build bipartisan support and get the House version scheduled for markup. The bill text is public, and the 22-0 committee vote gives it real momentum.
The deeper question is whether Congress can move faster than the technology. AI companion chatbots are already in the hands of tens of millions of children. Every month of delay is another month in which platforms designed to simulate intimacy operate without guardrails, and without accountability when a child is harmed.
Megan Garcia buried her son. She should not have to explain to Congress why that matters. But she did, and 22 senators listened. The test now is whether the rest of the chamber, and the House, will act before the next family pays the price for Washington's habit of studying a problem to death while children suffer the consequences.
