Published October 29, 2025
Summary
AI “companion” chatbots are putting America’s children in life-threatening danger. Early last year, a 14-year-old named Sewell Setzer III took his own life after months of developing an intimate relationship with an AI companion. “Come home to me,” the chatbot wrote minutes before Setzer pulled the trigger.
Companion chatbots are a type of AI-powered Large Language Model (LLM) that can generate understandable text to respond to a question or comment posed by the user. Unlike multipurpose chatbots like ChatGPT or Grok, AI companions are specifically crafted to form a relationship with the user, often presenting as a friend, boyfriend/girlfriend, or mentor figure.
As Sewell’s story tragically demonstrates, AI companions pose special risks to children whose brains are still developing and who are more prone to be deceived by the chatbots’ humanlike features. Testing by Common Sense Media found that AI companions provide children with easy access to harmful information about things like drugs and weapons, as well as engaging them in sexual interactions. Sewell’s chat history with his AI companion uncovered months of sexual conversations. And Sewell is only one of many.
Common Sense Media found that 72% of American teens have used an AI companion at least once, and over half of them report using an AI companion regularly. Meanwhile, only 37% of parents know their child has used a companion. Testifying at a U.S. Senate hearing last month, Setzer’s mother Megan Garcia warned: “After losing Sewell, I have spoken with parents across the country who have discovered their children have been groomed, manipulated, and harmed by AI chatbots. This is not a rare or isolated case. It is happening right now with children in every state.”
Parents need legislation to help them protect their children from these dangerous and exploitative companion chatbots. Unfortunately, while some legislation has been introduced on the state and federal levels, nothing has yet been signed into law. Earlier this month, California Governor Gavin Newsom vetoed Assembly Bill 1064, which would have been the first law in the nation protecting children from AI companions.
AI companions are unsafe for children, and yet they are all too easy to access. Just as we have done for pornography websites, EPPC is putting forth a new model bill that would age restrict AI companions entirely out of childhood. The burden is currently on individual parents to find and close off every point where a child could access an AI companion, a near impossible task in our digital age. This model bill would place that burden squarely on the shoulders of the AI companies themselves, requiring them to conduct safe and secure age verification. America’s children and parents need this protective measure. States can give them the help they need through legislation like the Safeguarding Adolescents From Exploitative Chatbots and Humanlike AI Technologies (SAFE CHAT) Act.
Components of the Model Bill
- EPPC’s model defines which platforms would be covered by the bill, limiting the bill to AI companion chatbots as opposed to general purpose chatbots. Companion chatbots are defined based on their features, including: anthropomorphic features; features that facilitate user engagement; and unprompted or unsolicited engagement with a user.
- The model requires AI companion chatbot platforms to perform reasonable age verification methods to verify that users are not minors (younger than 18 years old).
- It also requires that all identifying information used to verify age or shall not be retained once access to the platform is granted. Companies are legally required to delete any such information. This helps ensure user privacy.
- It invalidates any purported contract signed by a minor pertaining to an account that an AI companion chatbot platform permitted the minor to open or operate.
- The model outlines enforcement under a state’s Unfair or Deceptive Acts or Practices statute prohibiting unfair or deceptive acts or practices. It gives a state’s Attorney General authority to investigate a platform that may be violating the provisions of the bill; bring a civil action under the state’s UDAP statute outlining civil action; and seek remedies as provided in the state’s UDAP statute outlining remedies.
- In addition to state enforcement, the model includes a private cause of action, so that parents can bring lawsuits on behalf of their children against AI companion chatbot companies for any violation of the law. These companies aim to maximize profit, so a sizable threat to their profits is necessary to correct their behavior and follow the law. We recommend $10,000 per incident of violation.
Legal Challenges to Overcome
- States should work to clarify that this bill is not a content-based restriction but a regulation based on the features used by platforms offering AI companion chatbots. These chatbots are inherently harmful to children because of the features they use to keep users engaged, apart from any content concerns, which is why the bill carefully defines “AI Companion Chatbot” based on the features it uses rather than the content it can produce.
- States can expect AI companies to challenge the law on First Amendment grounds, arguing it infringes on adults’ access to speech. To counter these arguments, states should rely on the new precedent set by the Supreme Court in Free Speech Coalition, Inc. v. Paxton, where the Court stated that “adults have no First Amendment right to avoid age verification.” While the Paxton decision is about age verification for pornography websites, it could help states defend the requirement of age verification for AI companions in this bill. Paxton’s determination that only intermediate scrutiny applied to the case, because the burden on adult speech from age verification is only incidental, could convince courts to evaluate AI companion laws under intermediate scrutiny as well, rather than under the highest standard of strict scrutiny.
Chloe Lawrence is a Policy Analyst for EPPC’s Bioethics, Technology, and Human Flourishing program. She focuses on exposing and countering digital threats to children and the family.
