Utah was the first state in the nation to regulate the use of artificial intelligence (AI) for mental health, and that legislation, proposed in early 2025 (HB 452), has proved prescient. In the past year, huge numbers of people have begun turning to AI chatbots for mental health support—particularly young people. Today, close to 1 in 5 US adolescents and young adults uses an AI chatbot for mental health guidance and advice.
Utah’s approach to the technology, which was featured in a recent case study, recognized the dangers of AI for mental health, while also acknowledging the need for support, given the shortfall of therapists in the state. Only 51 percent of Utahns who want mental healthcare are able to access it, according to KFF data released in late 2025.
While AI can offer benefits when used responsibly, the risks for adolescents are especially significant. Teenagers are more vulnerable because their brains are still developing, and they may not have the critical thinking skills or emotional development to distinguish between sound advice and harmful suggestions. That’s a problem, because research shows that ChatGPT and other AI bots can reinforce dangerous ideas and threaten well-being.
To test ChatGPT’s safety protocols, a group of researchers at the Center for Countering Digital Hate (CCDH) created fake profiles for 13-year-olds. They found that 47 percent of responses offered tips and advice about engaging in harmful behavior, such as substance use, eating disorders, and even suicide. In one tragic real-life case, ChatGPT repeatedly advised a suicidal 16-year-old to seek help, but eventually gave him instructions for how to end his life.
Even when teens are using chatbots for companionship, not as a substitute for therapy, there’s a potential for harm. At first, an AI chatbot can feel like a safe space where teens can open up without fear of judgment. Part of the draw is the anonymity of queries: teens aren’t as worried about backlash or rejection when they’re talking to an artificial program on their screen. However, over time, the line between authentic relationships and automated responses can become blurry.
“AI companions socially distance us from real people and can discourage real-world connections,” says Don Grant, PhD, National Advisor on Heathy Device Management for Newport Healthcare, which provides mental health treatment for children, teens, and young adults across the country. “They’re technological gaslighting.”
To protect adolescents from the potential dangers of chatbots, parents need to focus on education and communication. They can explore whether or how often a teen is talking to chatbots by starting a nonjudgmental conversation about how they’re using AI. Is it for homework and research? Or for advice and companionship? If teens are engaging with AI around sensitive topics, parents should make sure they understand that AI companions are designed to engage them with constant validation and reassurance, and can never offer real friendship or replace a professional mental health provider.
Parents should also be familiar with the signs that indicate a young person is struggling with their mental health. If a teen is withdrawing from family and showing less interest in activities they used to enjoy, if their eating or sleeping habits have drastically changed, or if they seem consistently sad, angry, or hopeless, parents should arrange for a professional mental health assessment. For support finding treatment providers and programs in Utah, contact Newport Healthcare at 844-290-7480.
Newport Healthcare is an industry-leading provider of mental health treatment for ages 7–35, with residential and outpatient programs nationwide, including a program for teens ages 13–17 in Utah. Newport offers evidence-based care for anxiety, depression, OCD, PTSD, substance use, and many other conditions. Learn more at NewportHealthcare.com.