The British Affiliation for Counselling and Psychotherapy (BACP) is warning concerning the rising dangers of youngsters utilizing AI instruments resembling ChatGPT for psychological well being recommendation.
Its new survey revealed that greater than a 3rd (38%) of therapists working with underneath 18s have shoppers looking for psychological well being steering from AI platforms. And virtually one in 5 (19%) therapists reported kids receiving dangerous psychological well being recommendation.
Therapists have advised the BACP that some AI instruments are offering probably dangerous and deceptive info together with encouraging kids to self-diagnose circumstances resembling ADHD and OCD, reinforcing avoidance behaviours and mechanically validating their emotions no matter what they specific. There have additionally been tragic instances the place AI instruments have given dangerously misguided recommendation and even inspired suicide.* Therapists are additionally notably involved about AI’s incapacity to supply real-time assist or intervene in disaster conditions.
Ben Kay, Director at BACP, which is the most important skilled physique for counselling and psychotherapy within the UK and has greater than 70,000 members, mentioned:
“It’s alarming that kids are more and more having to show to AI chatbots like ChatGPT for psychological well being assist, usually unable to inform whether or not the recommendation they’re getting is secure and even true. Some have already suffered devastating penalties. And that is seemingly simply the tip of the iceberg, with many extra kids struggling in silence, with out entry to actual remedy.”
“We wish mother and father, carers, and younger folks to know that utilizing AI for psychological well being assist isn’t the straightforward, secure, or fast repair it would look like, there are actual dangers concerned, and it have to be approached with warning. Whereas AI is accessible and handy, it could’t replicate the empathy, connection, or the security of remedy delivered by an actual particular person educated to know complicated psychological well being challenges and assess dangers. Kids in misery may very well be left with out correct skilled assist. The data shared with AI additionally doesn’t have the identical protections as remedy. “
“Too many younger individuals are turning to AI as a result of they’ll’t get the psychological well being assist they want. That’s unacceptable. The federal government should step up and make investments now in actual, skilled remedy by way of the NHS, faculties, and neighborhood hubs. No younger particular person ought to ever be pressured to show to a chatbot for assist. AI would possibly fill gaps, however it could by no means change the human connection that adjustments lives. Younger folks deserve greater than algorithms, they deserve professionally educated therapists who hear.”
New survey findings
The BACP’s annual Mindometer survey, which gathered insights from almost 3,000 practising therapists throughout the UK, reveals that greater than 1 / 4 (28%) of therapists – working with each adults and kids – have had shoppers report unhelpful remedy steering from AI. And virtually two-thirds (64%) of therapists mentioned that public psychological well being has deteriorated since final yr, with 43% believing AI is contributing to that decline.
Senior accredited therapist Debbie Keenan who works at a secondary faculty and has her personal personal observe in Chepstow, added:
”I’m undoubtedly seeing extra kids and younger folks turning to AI to hunt remedy recommendation and self-diagnosis circumstances resembling ADHD and OCD. This raises actual issues for me. As superior as AI is, it merely can not do that. It additionally can not inform if a baby is distressed, dysregulated or at risk. If a baby was telling me that they had been going to harm themselves, or that they had suicidal ideation, assist can be in place for that baby earlier than they left my room- however would AI do that?
“Moreover, I’m additionally involved concerning the present danger of youngsters isolating and disconnecting from actual human relationships – this will result in an over reliance on Al for emotional assist and enhance emotions of loneliness, making it more durable to achieve out for ‘actual life’ assist.
“I consider kids are more and more turning to AI for remedy as a result of it’s obtainable 24/7. It feels non-judgemental and presents a way of privateness. Nevertheless, AI remembers information, it isn’t certain by moral or confidentiality requirements, and it lacks regulation or accountability. Whereas it might fill the hole in entry to psychological well being assist, it can not change human connection or recognise delicate emotional cues like a educated psychotherapist can.”
Amanda MacDonald, BACP registered therapist who offers assist for youngsters, teenagers and adults mentioned:
“AI remedy bots are inclined to undertake one among two approaches: providing validation or offering options. Each lack the nuance of actual remedy and danger giving recommendation that contradicts greatest practices for emotional misery. For instance, some AI instruments have suggested people with OCD to proceed their compulsions, mistaking short-term aid for progress. Others have inspired avoidance of tension triggers, which can really feel useful initially however can worsen nervousness over time by reinforcing avoidance behaviours.
“There have additionally been well-documented, tragic instances the place AI instruments have given dangerously misguided recommendation and even inspired suicide – outcomes which might be each devastating and deeply alarming.
“Dad and mom and carers ought to be conscious that their kids could also be turning to AI for steering and recommendation. Whereas it’s vital to maintain applicable parental controls in place, open and trustworthy communication at house is simply as very important. Discuss to your kids with curiosity and share your issues in an age-appropriate means.
“Kids and adolescents aren’t but outfitted to completely assess danger, so mother and father play an important function in protecting them secure. Balancing privateness with security is rarely straightforward, however with out that steadiness, younger folks can turn out to be overly reliant on what’s finally a really sensible algorithm; one which lacks the moral and safeguarding requirements present in helplines, remedy, or school-based assist.
“Reaching for his or her telephones once they’re upset feels pure for a lot of younger folks, particularly as AI instruments can appear supportive and validating. This creates a invaluable alternative for households to speak about their relationship with telephones and expertise. Dad and mom might help by modelling wholesome behaviour – setting shared screen-free occasions and recognising once they themselves instinctively flip to their telephones. In spite of everything, telephones had been designed to attach us, but when we’re not cautious, they’ll begin to change actual human connection.”
References:
All figures are from BACP’s annual Mindometer survey of its members. The overall pattern dimension was 2,980 therapists, and fieldwork was undertaken between 3 – 17 September 2025. The survey was carried out on-line.
