Survey finds most teen AI mental-health users still wanted more support

Teenagers turning to AI for mental-health support were usually seeking help elsewhere, too. A study published October 27 in npj Mental Health Research found that 92.1% of surveyed AI users had accessed other support. Yet most answering a separate question wanted additional help—a measure of unmet need, not proof that AI itself had failed them.

The survey began in classrooms

The findings come from the 2025 OxWell Student Survey, conducted during school hours between February and May that year. Researchers analyzed responses from 24,132 students attending secondary schools and further education colleges in England who reached the questions about general-purpose AI and large language models—the technology behind text-generating chatbots.

Schools were recruited through invitations from local authorities and education groups, rather than randomly selected. Most participating schools were in Oxfordshire and Liverpool. Those choices limit how far the results can be applied: this is a picture of the participating students, not a representative estimate for every adolescent in England, as News-Medical’s account explains.

In that sample, 662 students, or 2.7%, said they had used AI platforms for mental-health support during the previous year. The question concerned that specific use, not chatbot use generally. Researchers also asked about perceived helpfulness, other sources of support and whether students had wanted more help, allowing them to examine AI within a wider search for assistance.

Helpful did not mean all needs were met

Among the 580 users who rated their experience, 324 described AI support as helpful. Another 31.4% rated it neither helpful nor unhelpful, while 12.8% called it unhelpful. The separate question about additional support drew 649 responses, with 410 saying they would have liked more mental-health help during the previous year.

55.9%Rated AI support helpful

324 of 580 AI users who rated their experience described the support as helpful.

63.2%Wanted additional support

410 of 649 AI users answering a separate question reported wanting additional mental-health support.

The wider support picture was mixed. While 610 AI users had accessed help elsewhere, 46.1% had used formal services during the previous year. Fifty-two students reported AI use without any other support. Another 57 used AI alongside other online resources but reported no offline support. Together, those last two groups represented 16.5% of AI users.

Who turned to AI—and what remains untested

AI use was more common among students reporting higher depression and anxiety symptoms, previous self-harm, frequent loneliness or frequent food insecurity. Transgender and gender-diverse adolescents also had higher odds of use than boys and young men, the comparison group. Asian or Asian British students and students in other minoritized ethnic groups likewise had higher odds of using the platforms.

Ratings of helpfulness showed fewer differences between groups. Asian or Asian British adolescents were the only group significantly more likely to describe the support as helpful. Unmet support needs were more common among that group, students with higher depression and anxiety scores, and those reporting previous self-harm. Frequent loneliness was also associated with wanting more support.

Published in October 2026, the study describes experiences recorded in early 2025; use may have changed since then. Its central unanswered question is clinical: it did not assess whether AI advice was appropriate, safe or effective in improving mental health. Positive ratings therefore describe perceived help, not demonstrated health benefits—a distinction that remains important even when students welcome the support.