A significant minority of young people have been turning to AI chatbots to discuss everyday anxieties and concerns, according to two recent polls.
A survey published in JAMA Network Open, based on responses from 1,058 adolescents and young adults aged 12 to 21 in the United States, found that 13.1 per cent reported using generative AI chatbots for mental health advice. Pew Research Centre produced similar findings in 2025, with 12 per cent of teenagers reporting that they sought emotional support or advice from AI chatbots.
Advocates for the technology say AI offers a space to reflect, release frustration or seek discreet reassurance.
Digital safety specialists say this shift signals broader social dynamics rather than simple curiosity about new technology. It also hints at attitudinal fault lines between generations over AI.
For parents, concern may arise if they discover their child has engaged in emotional exchanges with AI. The instinct may be to restrict access or respond abruptly. Experts advise against sharply negative reactions.
Dr Melissa Morrison, an educational and child psychologist based in the UK’s West Midlands, says a parent’s initial response can shape whether a child feels safe enough to explain.
“If parents move in too harshly or judgementally, children will stop talking,” Morrison says.
A calmer approach, using open questions, can help parents determine whether a child was seeking reassurance, information or a private outlet, she explains.
Psychologist Ameena Zaheer says young people may be reluctant to discuss vulnerabilities with their parents because they fear they will overreact.
AI may seem an easier option because it removes the pressure of a human response. Acknowledging this dynamic does not mean encouraging reliance on technology, Zaheer says. Instead, it means recognising that the emotions driving the interaction are genuine.
Specialists also warn that the confident language used by AI can be mistaken for reliable advice.
Although chatbots often communicate with clarity and authority, experts stress that AI does not interpret emotional nuance or personal context in the way humans do.
Catharina Doria, an AI ethicist based in Brazil, says this distinction becomes critical when children seek guidance.
“AI systems generate language, not judgment,” she says.
Fluent answers delivered without hesitation can create the impression of understanding. Typically, AI produces responses by predicting probable word patterns rather than assessing what is appropriate or safe for a specific individual, Doria explains.
“The system does not assess vulnerability or risk,” she adds.
Encouraging children to question responses, verify information and discuss sensitive advice with trusted adults introduces an essential safeguard.
“Children need to understand that AI does not ‘know’ them,” she says.
Psychological safety is only one issue. Digital privacy is another layer of concern.
Cybersecurity specialists warn that many young users misunderstand how private AI conversations are. Exchanges may feel confidential, yet AI platforms do not function like discussions with a parent, teacher or clinician.
Andrey Sidenko, lead web content analyst at cybersecurity company Kaspersky, says information shared with AI systems is processed to generate responses and, depending on platform policies, elements of those interactions may be stored.
“In some cases, stored data may be accessible to third-party service providers involved in running or supporting the platform,” he says.
Any retained personal information is at risk if systems are compromised.
“The involvement of third parties increases potential exposure in the event of a security breach,” Sidenko says.
Parents should not be afraid to help their children understand these boundaries.
“Parental awareness and involvement is crucial,” he says.
Clear, age-appropriate conversations about what AI is, how it works and what should remain private, can turn uncertainty into understanding, he adds.
Nour Ibrahim is a 2025-26 Rosalynn Carter Mental Health Journalism Fellow.
