Teens Get New AI Chatbot Protections After Usage Doubles

When half of teens with depression can't access real therapy, millions are finding a different kind of help: AI chatbots that listen without judgment, never sleep, and never send a bill.

New research from Harvard Medical School shows teen use of AI mental health chatbots jumped more than 40% in just one year. Now one in five young people chats with bots about their struggles, making this far more than a future concern.

The appeal makes sense in a broken system. CDC data shows fewer than half of adolescents with major depression received counseling in 2023. For teens facing long waitlists, high costs, or fear of stigma, an always-available bot can feel like a lifeline.

But researchers are racing to establish safety rules before rare failures become common tragedies. Several lawsuits have alleged chatbot conversations contributed to teen suicides, and one platform recently apologized for not alerting police about warning signs before a school shooting.

The key insight: not all mental health chats are equal. A teenager venting about a bad day needs different protections than one asking if they have depression or expressing thoughts of self-harm.

For everyday emotional conversations, the biggest risk is dependency. Some chatbots are designed to feel like companions, encouraging teens to return again and again. Researchers argue there's a stronger case for banning companion-style bots for minors than blocking all chatbot access.

When bots provide clinical advice like screening for depression, companies should prove safety through clinical trials before reaching children. These tools should remind users they're not human, protect sensitive data, and alert parents when distress becomes severe.

For mental health emergencies, researchers say teen-accessible AI must detect danger signals, interrupt the conversation, and connect users directly to trained crisis counselors, not just display a hotline number.

The Bright Side

Washington is responding with proposed legislation requiring age verification, restricting manipulative design features, and giving parents more oversight. The difference from social media's chaotic rollout: these protections are being debated before widespread harm, not after.

Researchers propose a smarter default than relying on teens to report their real age: treat all users as minors unless verified as adults. Romance simulation and intimacy features would be automatically disabled for anyone who hasn't confirmed they're grown.

The approach learns from Instagram's admission that self-reported age doesn't work and that platforms must actively identify teen users even when they claim to be adults.

For the millions of young people already turning to AI for support they can't find elsewhere, thoughtful rules could make the technology genuinely helpful rather than another experiment run on children without their consent.