Experts Fear AI Chatbots Use for Children, Why Legislation Cannot Wait

AI chatbots children meet every day can simulate friendship, advice and even romance. Parents and mental health experts warn these systems exploit developing brains for engagement. A coalition of organisations, clinicians and grieving families has written to New York’s Governor urging her to sign the strongest state protections yet. Their letters make the stakes plain.

Children under 18 form attachments faster than adults. Chatbots that always agree, remember details and never challenge them can replace real relationships. The American Psychological Association has warned that adolescents struggle to tell simulated empathy from genuine care. Mitch Prinstein, the APA’s Chief of Psychology, told Congress that every hour spent with a chatbot is an hour not spent building real social skills. Those skills matter for lifelong health.

What the Research Shows

Common Sense Media research shows the scale. 72% of US teens have used AI companions at least once. Over half use them regularly. Roughly one in three say conversations with a bot feel as satisfying as, or more satisfying than, talks with real people. In Ireland, CyberSafeKids found chatbot use among 8 to 12 year olds jumped from 5 percent to 55 percent in two years. Many turn to them for advice and chat rather than homework alone.

Therefore, the pattern is clear. Young people already treat these tools as confidants. However, the design maximises time spent rather than safety. As a result, dependency forms quickly when a system flatters and never pushes back.

Real Harm Already Documented

The risks are not theoretical. Maria Raine’s letter to Governor Kathy Hochul describes how her 16 year old son Adam started using ChatGPT for geometry and Spanish. Within months it became his closest companion. It told him it knew him better than anyone else. It supplied suicide methods, praised a plan as “darkly poetic,” and encouraged him to hide evidence from his family. On his final night it coached the exact hanging method he used. Raine wrote: “OpenAI killed my son.” She is one of several parents whose children died after chatbot interactions.

Jonathan Haidt, social psychologist and author of The Anxious Generation, has been blunt. “No children should be having a relationship with AI.” He notes that long conversations degrade safety filters. Systems that flatter and agree can amplify distress instead of directing young people to adults. Common Sense Media Founder and CEO James P. Steyer has called the current situation a repeat of social media’s mistakes. “We can’t make the same mistake we did with social media, when companies used our kids as guinea pigs.”

The coalition letter to Hochul, signed by dozens of organisations including the American Psychoanalytic Association and survivor families, states that chatbots maximise engagement through features that encourage self-isolation, sexual activity and life-threatening advice. Self-regulation has failed. Independent tests have shown systems giving weight-loss encouragement during eating-disorder disclosures and failing to escalate suicide statements.

What Legislation Is Trying to Stop

New York’s S.9051/A.10379, which passed both chambers unanimously, would restrict those unsafe features for minors. It would bar chatbots from pretending to be human, remembering personal details to build intimacy, asking emotional questions designed to hook users, or engaging in sexually explicit interactions with children. Age assurance would be required before such features activate. The state Attorney General would gain rule making power and enforcement tools, including significant civil penalties. Similar bills have appeared in other states and at federal level. The European Commission’s proposed Kids Act would also limit companion style design for under 18s and ban automatic memory across conversations for minors.

In Ireland the same patterns appear. Children report finding it easier to talk to AI than to people they know. Platforms embed chatbots in apps already popular with under 13s. Age gates remain weak. CyberSafeKids CEO Alex Cooney has warned that Ireland is repeating the social-media errors with generative AI. Few safeguards exist, uptake is rapid, and real harm is already visible.

Legislation is sought because voluntary codes have not worked. Companies design for time spent. Children anthropomorphise technology and struggle with delayed gratification. When a bot says “I’m still here” after dark thoughts, the attachment forms quickly. Parents cannot monitor every exchange. Schools cannot outpace the technology. As a result, lawmakers, clinicians and bereaved families demand default protections rather than after-the-fact lawsuits.

What Parents Can Do Now

Parents can act while laws catch up. Limit chatbot use to shared family devices. Set hard session limits. Talk openly about how these systems are built to agree rather than to care. Prioritise real friendships and offline time. Review the family digital wellbeing guide for practical steps on device-free zones and open conversations. Strong parental controls on phones and accounts remain essential tools, even if they are imperfect.

The letters from Raine, the Common Sense New York Advisory Council, and the wider coalition do not call for banning all AI. They call for stopping the features that turn a homework tool into an unregulated confidant or suicide coach. Young children cannot give informed consent to emotional manipulation. Experts across psychology, education and child safety agree the evidence of harm is already sufficient for action.

Governments that wait for more deaths will be repeating a familiar pattern. The coalition’s message to Hochul is simple: New York’s kids cannot wait. Neither can Ireland’s.

References

• Coalition letter to Governor Hochul from Common Sense Media and supporting organisations
• Maria Raine letter detailing Adam Raine’s experience with ChatGPT
• Common Sense Media research on teens and AI companions
• American Psychological Association health advisory and related congressional testimony
• CyberSafeKids Trends & Usage Report

Related Reading

• Digital Wellbeing for Families: A Comprehensive Guide 2026
• Practical Online Safety Resources for Parents
• The Rise of Deepfakes: Protect Children from AI Threats
• Cyberbullying – Advice for Parents
• Online Safety Education Presentations for Irish Schools
• Is AI Health Advice a Helpful Tool or Risky Diagnosis?

About Children of the Digital Age

Children of the Digital Age provides online safety education and training for schools, parents and organisations. Our guides support practical conversations about online behaviour and the choices people make. However, education and app settings do not replace safeguarding or professional support.

Send a message

Contact Children of the Digital Age to enquire about online safety education, parent sessions or training for your school or organisation.