Top Takeaways
- Teens are often aware of the potential harms of using AI for mental health support.
- Experts say the future likely involves a combination of in-person care and AI chatbots.
- Gov. Gavin Newsom signed into law landmark protections around teens’ use of AI chatbots on Thursday.
How can I make sense of my father’s absence? How should I respond to hostility on social media? And during a period of tension, how can I make myself feel at peace?
These are some of the questions teens told EdSource they have asked AI chatbots in recent years, turning to the technology for emotional support and advice. They’re not alone — roughly 1 in 5 teens now use AI chatbots for mental health advice amid an adolescent mental health crisis exacerbated by limited access to professional care.
EdSource this week reported on the rise of teens’ use of AI for mental health support, and how some schools have embraced using customizable AI chatbots to flag kids in crisis. Here are three takeaways from that reporting.
Teens recognize potential harms
Teenagers are aware of the potential downsides of turning to AI for mental health support: It’s designed to affirm their feelings and can’t always be taken at face value.
“Everywhere that AI extracts information from, not all of it is credible,” said Milan Li, a senior at Taft Charter High School in Woodland Hills, a Los Angeles neighborhood. “And especially with these very heavy matters, I don’t think it’s always a good idea just to search up an AI prompt and follow exactly whatever they tell you to the T because, more often than not, it’s not suitable for you.”
Still, teens continue to turn to AI when they need emotional support or advice. Vihaan Bhardwaj, a senior at Mt. Everest Academy and the K-12 co-chief of policy at the GenerationUp youth advocacy group, said he’s turned to AI chatbots for emotional support or advice on several occasions, including during social media conflicts with peers.
Bhardwaj said he eventually realized that the chatbots weren’t supporting him in the ways he needed — and that a therapist would have been better equipped to guide him.
“It backfired,” Bhardwaj said, “because I feel like I didn’t get to process that emotion fully … You feel this incompleteness, and the emotions still linger.”
Another teenager, Liliana Messinger, decided to change her relationship with AI entirely after relying on it for nearly a decade.
The future comes with a blend
Much of the debate over AI and mental health tends to fall into extremes: Either chatbots will cost lives or can help save lives. The reality is likely somewhere in between, with AI potentially supplementing — but not replacing — human support, according to experts.
Stephen Aguilar, a professor of education at USC, said companies developing chatbots for mental health should work closely with mental health professionals and build in safeguards that trigger additional support when “a student crosses some sort of red line.”
Alongside, a Seattle-based AI wellness company, has developed a chatbot for mental health support that flags concerns of self harm to a trusted adult. More than 200 school districts nationwide have contracted with the company to deploy AI chatbots for students.
Experts told EdSource that AI chatbots could help fill in gaps in mental health care, particularly for those who can’t access or afford it.
But for that to happen, Aguilar said, counselors and school staff need to teach students how to use chatbots responsibly.
“There are no guidelines, and students use one thing, and then adults sort of see that they’re using it and are reactive to it,” Aguilar said, “as opposed to being well-versed in what’s there to begin with, and then helping teens have a better understanding of what those limitations are.”
More regulation is on the way
As more children and teens turn to AI chatbots for mental health support, safety experts are concerned about emotional dependence, manipulation, parasocial relationships and social isolation.
A recent risk assessment by Common Sense Media’s Youth AI Safety Institute found that many consumer-facing apps posed “unacceptable risk” to teens, with serious failures involving sexual content, eating disorders, psychosis and suicide risk.
Robbie Torney, senior director of AI Programs at Common Sense Media, said regulators need to “hold tech companies accountable” to make sure chatbots meet “minimum safety standards” before they are used by minors. He also said Congress should establish clear liabilities for platforms that harm children, require mandatory safety testing and allow states to develop their own AI policies to safeguard children and teens.
Two bipartisan proposals are before Congress — the SAFE KIDS Act and the CHAT Act. 2.0 — which would require AI systems to prevent emotional manipulation, require parental notification in situations of self-harm and stop AI companions from impersonating mental health professionals to minors. But the Trump administration has broadly opposed safeguards, issuing a December 2025 executive order aimed at challenging state AI laws.
California has led the nation in regulating AI. On Thursday, Gov. Gavin Newsom signed into law sweeping legislation that requires protections around companion chatbots for children, including crisis protocols and parental controls, and mandates annual child safety audits.
“It shouldn’t take kids losing their lives before we take action,” said Keri Rodrigues, president of the National Parents Union.
‘We actually saved a kid’: Schools recruit AI chatbots as counselor shortage persistsSeptember 10, 2026
Teens use AI for academic support even though they see the dangers, survey findsAugust 18, 2026
This story was originally published by EdSource.