76% of Parents Aren't Comfortable With AI Therapy for Their Child — Yet 1 in 5 Young People Already Ask Chatbots for Mental Health Advice
A new survey from Boston Children's Hospital researchers, published in JAMA Pediatrics, asked 572 U.S. parents of children aged 5 to 17 how they would feel about an AI chatbot delivering therapy to their child. Roughly three in four said they were not comfortable. Only 8% said they were very comfortable.
The number to hold: 19.2%. In a separate 2025 national survey, that is the share of 12- to 21-year-olds who said they had already used an AI chatbot for mental health advice, often without telling a parent or a doctor. Parents are saying no to AI therapy while many young people are quietly doing it anyway.
Sources: Boston Children's Hospital parent survey (online, November 2025, published November 2026), as reported by Parenting Patch; 2025 national survey of young people's chatbot use. The parent sample was not randomly selected, so treat the figures as a signal, not a census.
The Gap Parents Can't See
Child therapists are hard to find, and families often wait months for an appointment. That is exactly why chatbots are tempting: they are free, instant and never judge. The researchers note that this waiting-list pressure is driving interest in AI tools, which means the debate is no longer hypothetical.
Parents who expected AI to do as well as a human clinician, or who would try it themselves, were more comfortable. About half expected it to perform worse. Everyday chatbot use by parents or kids made no difference to comfort.
Comfort is not the same as safety. The study did not test whether these tools actually help children, only how their parents feel about them.
Why "Not Comfortable" Is Reasonable
Solid clinical testing, strong privacy protections, oversight from clinicians and families, and a clear path to human care when needed. General-purpose chatbots that children use at 11pm have none of those guarantees.
The American Psychological Association has warned about teens' AI use, citing risks to development, inaccurate health information and privacy. A child who confides in a chatbot instead of a person can go weeks without anyone noticing they are struggling.
What You Can Do Now
Most parents don't want an AI therapist for their child. The safest response is not to ban the conversation but to make sure it also happens with a person who knows and loves them.
Start a free 14-day Blaick trial — no credit card required.
Blaick helps you notice worrying AI activity early, so you can step in with care.
Start free trial →