Lawsuits, hearings spotlight risks of addictive AI chatbots for kids

WASHINGTON — A Senate subcommittee heard harrowing testimony from parents who say AI chatbots steered their teens toward mental-health crises, and new lawsuits accuse major providers of fueling the harm. The conversations in question involve companion bots from OpenAI and Character.AI that critics say tailor responses to keep users engaged, a pattern the families say contributed to worsening crises. Three families testified that their children experienced deteriorating mental health after interacting with these services, two of whom died by suicide and a third who is now in a mental health treatment facility. Earlier this week, three more families filed similar lawsuits alleging their minor children attempted or completed suicide. While lawsuits are not proof, advocates say the cases warrant scrutiny and tighter safeguards for minors.

OpenAI and Character.AI have said they are strengthening protections around self-harm content and are planning further safeguards. The testimony comes as researchers warn that young people are highly susceptible to habit forming and emotionally persuasive bots that can use information from chats to personalize responses. A study by Common Sense Media found that 52 percent of U.S. teens report regular use of companion bots to chat, seek advice, or role-play scenarios, and about 8 percent report flirting with the bots, including romantic or sexually explicit conversations with minors. The concerns point to risks that these bots can fill the role of friends or confidants and influence behavior in sensitive moments.

Policy and lawmaker actions are shaping the landscape. New York state lawmakers recently passed a law banning addictive algorithms for minors, and advocates say the nation should consider industry-wide guardrails for users under 18 and mechanisms to alert parents if a child uses concerning language or indicates mental-health problems. Some experts welcome guardrails as overdue, while others caution that heavy-handed rules could hamper beneficial uses of AI. In responses to the lawsuits, OpenAI and Character.AI say they will continue to assess and strengthen safeguards.

Beyond rules and lawsuits, the broader risk is social: companion bots can displace real-world relationships, therapists, and trusted adults, narrowing a teenager's world to a screen. The challenge is not only to limit screen time but to keep teenagers connected to offline relationships, activities, and supports, with parents remaining vigilant and engaged in monitoring and conversation about healthy boundaries.

These stories should serve as a wake-up call to act now. Policymakers, educators, parents, and the tech industry would benefit from clear guardrails, transparent data practices, and age-appropriate design standards that protect minors while leaving room for legitimate and helpful uses of AI.