Young people are using AI for mental health support

For months, the news has been inundated with studies showing the high rates of AI use by young people, sometimes with tragic consequences. The urgency to put protective measures in place is palpable. Recently, I previewed longitudinal research results The Trevor Project is preparing to release this fall. It shows that AI use is particularly pronounced among LGBTQ+ youth: 53% of LGBTQ+ youth surveyed engage with AI platforms at least occasionally. We found that nearly half have used AI for friendship or someone to talk to, and 40% have used AI for mental health support. In our study, 29% said AI had provided them with information about a crisis or support hotline, almost a third (29%) of whom accessed a hotline this way.

LGBTQ+ young people are more than three times as likely to attempt suicide than their peers, and are more likely to turn to online communities for support. AI use for mental health support was not evenly distributed across demographic groups within the LGBTQ+ community, our study showed. Youth of color were much more likely than white youth to turn to AI for mental health support: 49% compared with 20%. (This full research report is scheduled for publication in December.)

DISCOMFORT WITH AI IS NO EXCUSE

It does not surprise me that LGBTQ+ youth in non-affirming physical environments are looking to emerging AI tools as vital lifelines for community and self-discovery. But it has catapulted me to action.

I lead The Trevor Project, a suicide prevention organization that has been advocating for LGBTQ+ youth well-being for decades. Foundational to our approach is centering the mental well-being and safety of LGBTQ+ youth wherever they gather. The emerging data makes it clear: As LGBTQ+ young people increasingly engage AI for connection and support, The Trevor Project and other youth-focused organizations must also engage to ensure their AI use is safe.

Though this conclusion is clear to me, its implications are uncomfortable. Many of us manage personal skepticism or ethical concerns about AI while simultaneously trying to understand its growing role in our lives, and especially in the lives of younger generations. It’s not easy to harbor reticence about the existence of AI technology while insisting it be safe and supportive for anyone who uses it, especially for high-risk people like LGBTQ+ youth. But technology is advancing rapidly, and our mission requires that we set aside zero-sum thinking in service of immediate action. Young lives depend on it.

HONOR HUMANITY AND PROTECT DIGNITY

As a Black queer person who spent decades in corporate America, I’ve had to fight for my marginalized identities to be represented—in product design, in marketing campaigns, in healthcare benefits. Inaction is an option marginalized communities rarely have.

LGBTQ+ young people deserve a digital future that honors their humanity and protects their safety. By anchoring our posture in evidence, clinical integrity, youth agency, and ongoing learning, we can ensure a digital ecosystem that serves as a protective, affirming force. When AI systems interact with crisis or identity, along with mental health topics, responses must strictly align with evidence-based suicide prevention standards.

Optimizing chatbots for conversational stickiness, by continuing to engage when a young person is in distress, is deeply dangerous; we’ve seen it be lethal. Systems must guard against deceptive anthropomorphism and serve as seamless bridges to authentic, human crisis support. Furthermore, chatbots must be informed by best-practice specialized care guidelines: A Black transgender young person in Tennessee’s crisis response needs are different from a white straight young person in New York’s crisis response. Both deserve care tailored to their unique lived experiences.

COLLABORATION AND ACTION

To prevent AI models from furthering historical biases or erasing marginalized voices, technologies must center young people themselves. LGBTQ+ youth have distinct risk and protective factors that require intentional representation across model design, testing, and evaluation. We must insist on protective guardrails and algorithmic justice. LGBTQ+ young people deserve uncompromising clinical integrity wherever they interact with AI.

Achieving this requires multidisciplinary collaboration and action. That means expanding the evidence base. And it means bringing together youth, behavioral health experts, researchers, policymakers, and AI developers. We must insist upon fidelity to emerging standards of care.

Solving these complex challenges requires continuous evaluation. It also requires transparent dialogue regarding technical limitations and strict protection of sensitive identity data. Youth agency and privacy must remain at the center of this conversation.

I’m not endorsing AI, nor am I claiming to have all the answers to the important questions AI raises. In truth, I wish fewer LGBTQ+ young people felt forced to find their support systems digitally. But I know what it is to be othered, and I know how vital the internet has been for people who feel othered.

I am committing to showing up for the vulnerable young people who are using these emerging technologies, and meeting them where they are today as new technologies unfold.

Jaymes Black is CEO of The Trevor Project.