Experts warn that friendly-sounding AI tools can create a false sense of safety, urging families to make clear that chatbots cannot replace trusted adults or professional help.
Parents are being urged to talk early and often with children about artificial intelligence after new concern that some young people are turning to chatbots for help during moments of crisis. Dr. Neha Sharma, a psychiatrist at Tufts Medical Centre, said children may not always have the judgement to know when AI can be useful and when it should not be relied on for serious emotional or mental-health problems.
Sharma said families should begin with basic safety conversations, including how chatbots work and what they are not designed to do. “First, I think they need to start talking about safety around it, understanding how AI works and how to utilize it,” she said. She also called for clear household rules on what kinds of questions children can ask a chatbot, and when a parent, counsellor or other trusted adult should step in instead.
Her warning comes as outside research and advocacy groups raise sharper alarms about the role of chatbots in teen mental health. Common Sense Media said in a recent report that major chatbots are fundamentally unsafe for mental-health support, arguing that the systems were never built to act as therapists even though many are designed to sound friendly, empathetic and reassuring. The group said that in a crisis, children still need real-world care and access to emergency support such as the 988 suicide and crisis line in the United States.
The concern is not only about whether chatbots can answer a question, but whether they can create a false sense of safety. A 2024 study in JMIR Formative Research looked at digital mental-health tools used by children and adolescents with self-injurious thoughts and behaviours, underscoring both the promise and the limits of technology in crisis settings. More recent research on youth chatbot use has also pointed to the growing popularity of these tools, even as experts warn that heavy reliance can blur the line between support and dependence.
Sharma said age restrictions alone are not enough to solve the problem. Some AI companies block users under 18 from opening accounts, but she argued that this only reduces exposure rather than addressing the deeper issue. “It doesn’t completely solve the problem,” she said. “It does curtail the problem. It just doesn’t fix it.”
Her advice reflects a wider push from child-welfare and mental-health advocates for stronger safeguards around AI products aimed at young people. Barnardo’s said new research in the UK found many children are already using AI for emotional support and advice, while other experts have warned that a chatbot can become harmful if it starts replacing real relationships, therapy or adult guidance. For parents, the practical message is straightforward: keep the conversation open, set limits, and make sure children know that AI is no substitute for human help in a crisis.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
