Study Examines Teen Use of AI Chatbots for Mental Health Support

A recent study published in the journal npj Mental Health Research has examined how adolescents in England turn to general-purpose artificial intelligence and large language model chatbots for mental health advice. Conducted by researchers including H. Okechukwu, H. A. Bear, M. Fazel, and E. Soneson, the investigation looked at data from more than 24,000 students to understand who uses these digital tools, their perceived helpfulness, and the presence of unmet support needs.

Survey Findings on AI Platform Use and Perceived Helpfulness

The researchers analyzed responses from the 2025 OxWell Student Survey, which gathered data from secondary schools and further education colleges in England primarily located in Oxfordshire and Liverpool. Out of 24,132 participating adolescents with a mean age of 14.0 years, 662 respondents—representing 2.7% of the total—reported using general-purpose artificial intelligence or large language model platforms for mental health support during the previous year.

Among the participants who rated their experiences with these tools, 55.9% stated they found the support helpful. Meanwhile, 31.4% described the advice as neither helpful nor unhelpful, and 12.8% found it unhelpful. However, the study also revealed that nearly two-thirds of AI users—specifically 63.2% of those who answered a separate inquiry—indicated they would have liked additional mental health assistance over the prior year. The study authors noted that this metric measured perceived unmet need rather than proving the AI platform itself was inadequate.

Context of Existing Support Systems and Vulnerabilities

The vast majority of adolescents who utilized AI platforms were also engaging with other forms of support. Overall, 92.1% of AI users reported receiving help elsewhere, and 46.1% had accessed formal mental health services during the previous year. Only 52 students reported using AI tools without any other support, while another 57 relied on AI alongside other online resources without offline backing. Combined, these two groups accounted for 16.5% of the platform’s adolescent users.

The analysis identified several demographic and social characteristics associated with a higher likelihood of using AI for mental health advice:

• Transgender and gender-diverse adolescents showed higher odds of platform use compared to boys and young men.
• Adolescents from Asian or Asian British and other minoritized ethnic groups reported higher odds of use.
• Higher usage was observed among individuals experiencing frequent food insecurity, elevated depression and anxiety symptoms, previous self-harm, frequent loneliness, or occasional bullying.

Study Limitations and Important Considerations

The study authors acknowledged several methodological limitations. Because school recruitment was nonprobabilistic and concentrated in specific regions of England, the findings may not represent adolescents nationwide. Furthermore, the cross-sectional design demonstrates associations but does not establish causation. The research did not formally evaluate whether the advice provided by general-purpose AI was clinically appropriate, safe, or effective in improving mental health outcomes. Experts emphasize that while digital tools are increasingly integrated into how young people seek guidance, they cannot replace comprehensive clinical care or assume the ability to meet complex adolescent needs.