The OpenAI Foundation is funding a new, one-year research initiative by the Child Mind Institute to study how young people interact with artificial intelligence chatbots. The study aims to map out how digital tools affect youth mental health and establish safety standards for future AI designs. This effort comes as more teenagers and young adults turn to general-purpose chatbots and digital companions for emotional support.
What Happened
The Child Mind Institute is running an independent study on how AI chatbots affect youth mental health. The OpenAI Foundation provides the funding, but the institute is doing the research on its own. Researchers will track young people's mental states using clinical assessments, digital journals, and anonymous chat histories. Rather than testing one specific app, the project focuses on building the infrastructure needed to design safer conversational AI tools for children.
The Bigger Picture
Demand for these tools is high. A study in JAMA Pediatrics found that nearly 1 in 5 young people aged 12 to 21, representing about 8 million youth nationwide, have used an AI chatbot for mental health advice. More than 40% of these users seek emotional support from chatbots at least once a month.
This trend points to a shortage of human therapist options. According to Science News, roughly 40% of adolescents who had a major depressive episode did not receive professional treatment. But general AI is not ready to help. A 2025 study of over two dozen chatbots showed that not a single chatbot gave an adequate response to a suicide risk query.
Experts warn against the overinflation of clinical claims made by tech firms. As seen in hands-on healthcare training, digital tools should support human experts instead of replacing them.
What This Means for Families
Many teenagers hide this behavior from adults. More than half of the young people who use AI chatbots for mental health advice keep it a secret. This leaves them vulnerable to privacy and cybersecurity risks, because these platforms store sensitive personal disclosures.
Young users are aware of these issues. In a study published in npj Digital Medicine, teenagers who helped co-design an AI mental health app demanded transparency, stating that they need to know what is under the hood of the technology. They wanted to "humanise AI without dehumanising care," preferring tools that assist rather than replace therapists.
What You Can Do
Parents and educators can start by asking teenagers directly if they have ever used chatbots like ChatGPT for advice when feeling stressed or anxious. It helps to set clear boundaries, explaining that while these tools can assist with schoolwork, they lack the programming to offer safe mental health or crisis advice. Finally, make sure young people have direct access to human-led resources, like the 988 Suicide & Crisis Lifeline, instead of relying on AI algorithms.