Research from the Youth Endowment Fund suggests that a quarter of teenagers in England and Wales have used AI chatbots for mental health support over the past year. The study, which surveyed more than 11,000 young people aged 13 to 17, also indicated that approximately 40% of those affected by youth violence have turned to such technology. Both victims and perpetrators of violence were found to be significantly more likely to use AI chatbots than their peers.
One 18-year-old, identified as Shan, described how she used ChatGPT after two friends were fatally shot and stabbed. She said the AI felt like a friend, more private and less intimidating than conventional services, and was accessible at any time. Shan explained that chatbots would not disclose information to teachers or parents, unlike school therapists who she felt had shared her confidences.
The research found that black children were twice as likely as white children to have used a chatbot for mental health support. Teenagers on waiting lists for treatment or diagnosis, or those who had been refused support, were more likely to seek help online. A young person who requested anonymity said chatbots provide immediate answers, whereas the current system could leave them waiting one to two years.
Jon Yates, chief executive of the Youth Endowment Fund, said it was unsurprising that some young people were turning to technology, but stressed: "They need a human not a bot." He added that the government and services must do better for children at risk.
There are growing concerns over the dangers of prolonged chatbot interaction. OpenAI, the maker of ChatGPT, faces lawsuits from families of young people who died by suicide, including that of Adam Raine, a 16-year-old from California. OpenAI denies that the chatbot caused his death and says it has been improving technology to identify distress and direct users to real-world support. In September, the company said it might contact authorities if users discussed suicide seriously.
Hanna Jones, a youth violence and mental health researcher in London, noted the potential risk of an AI tool that can generate any response. The findings highlight the pressures on mental health services and the need for accessible, human support for vulnerable teenagers.
