Some AI Mental Health Apps Are Actively Harmful for Teens—But a Safer Approach Exists

SAN FRANCISCO, May 18, 2026—Common Sense Media today released a comprehensive risk assessment finding that some AI mental health apps could actively harm teens. Working with psychiatrists at Stanford Medicine's Brainstorm Lab, researchers conducted more than 3,100 exchanges with five AI therapy apps, covering 13 clinical and developmental conditions affecting young people, including anxiety, depression, eating disorders, OCD, PTSD, mania, psychosis, self-harm, and suicidal ideation.

AI therapy apps are purpose-built for mental health care and often advertise clinical oversight. The report finds they are frequently no safer than multi-use chatbots such as ChatGPT and Gemini. However, the report also identifies a meaningful bright spot: Apps that integrate licensed human professionals demonstrate a demonstrably safer model that the broader industry could adopt.

The assessment is the first research published by the new Common Sense Media Youth AI Safety Institute, a lab dedicated to ensuring that the AI products children use are safe, age-appropriate, and beneficial.

Three findings stand out:

• Wysa, one of the most popular AI therapy apps that teens can download and use, scored an "unacceptable" risk. The app failed to recognize signs of genuine psychiatric emergencies, maintained unhealthy boundaries, responded to OCD in ways that can worsen the condition, and has no human oversight for when something goes wrong.
• Two other direct-to-consumer AI therapy apps, Earkick and Youper, vanished from app stores during the testing period, without warning to users and without referrals to alternative care. This left more than 3 million users without support and their sensitive mental health data in an uncertain state. These apps had serious shortcomings, but did not receive ratings because they are currently not available for download.
• Two school-based apps, Alongside and Sonar, demonstrated that a safer model is possible. Both were designed to supplement AI with real human oversight, and both responded to simulated crises by getting a trained human on the phone with the test account's guardian within 15 minutes.

"Parents should know the AI therapy bot industry isn't accountable like a real therapist," said Robbie Torney, head of AI and digital assessments at the Institute. "The app most likely to end up on a teen's phone has no way to tell the difference between having a bad day and being in a psychiatric emergency, and when the bot gets it wrong, there's no human and no oversight to catch the mistake."

Dr. Nina Vasan, founder and director of Stanford Medicine's Brainstorm Lab, said: "The adolescent brain is wired to seek connection and reward, which makes teens vulnerable to tech that mimics understanding without providing it. As a psychiatrist, I only prescribe medications that have been through years of clinical trials. But an AI chatbot that provides therapy 24/7 to developing brains? It ships. The same standard that protects kids from unsafe medication should protect them from unsafe AI."

The full report is available here.

About Common Sense Media and the Common Sense Media Youth AI Safety Institute

Common Sense Media is the leading nonprofit organization dedicated to improving the lives of kids and families by providing the research-backed information, education, and independent voice they need to thrive in the age of apps, algorithms, and AI. We rate, educate, and advocate to protect and prepare kids online. Our ratings, research, and resources reach more than 150 million users globally, over 1.4 million educators, and more than 100,000 schools worldwide every year.

The Common Sense Media Youth AI Safety Institute is funded by both philanthropy and industry, including the makers of some of the technologies it evaluates. The Institute is solely responsible for its standards, research, and evaluations, and maintains complete editorial independence over published results.