In the U.S., more people are seeking mental health advice from artificial intelligence tools, and therapists have mixed feelings about this trend.
A recent study by the American Psychological Association revealed that 39% of psychologists reported discussions with patients who had turned to AI for self-diagnosis. Roughly one-third of the professionals indicated that their patients depend on AI for support with self-control, affirmations, and reminders, while the same percentage acknowledged its use to enhance therapy.
Experts from the Daily Caller News Foundation highlighted that while AI offers an accessible and budget-friendly way to obtain mental health information, it can also introduce challenges. For instance, patients may encounter biased information.
Adam O’Malley, a researcher with the Cato Institute, noted that AI is appealing because it is cost-effective, well-informed, personalized, and engaging. “It’s understandable why people would seek out AI for mental health support—it’s able to simplify complex research and provide a sense of companionship,” he explained.
He pointed out that AI utilizes some of the best psychological research; it can be tailored to individual preferences and is knowledgeable about nearly any subject matter or cultural context. However, he cautioned that AI could be influenced by biased training data. Hence, selecting the right information for training is crucial, especially since much psychological research remains unverified.
O’Malley emphasized that traditional face-to-face therapy won’t disappear anytime soon. “Even if AI-driven therapy is not as effective, its affordability and accessibility could make it a beneficial addition to the mental health care landscape,” he stated.
AI tools are also useful for administrative tasks in mental health care, such as managing appointments, facilitating communication, and summarizing health records.
Psychologist Dr. Zachary Ginder from California remarked that the convenience and low cost of AI are driving many people to seek mental health information through these tools. “It’s low-effort and readily available from personal devices, which feels private and is inexpensive,” he shared. “This allows those needing quick help with mental health issues to avoid the complications of insurance or the stigma that often comes with professional help. But of course, there are big drawbacks and risks.”
He advised against relying on large language models (LLMs) like ChatGPT or Claude for specific clinical treatments or counseling, noting that many consumer chatbots marketed as mental health solutions lack clinical validation.
Chatbots, which can mimic human conversations, are often powered by AI. Some studies indicate that these chatbots may show more bias towards certain mental health issues, like alcoholism or schizophrenia, in comparison to others, such as depression.
Ginder further elaborated on the absence of consistent regulation in the U.S., mentioning that these tools often operate outside any oversight framework, raising concerns about serious repercussions, including tragic incidents involving AI and suicide.
Recently, some states have passed laws restricting AI chatbots from giving mental health advice to younger users, following reports of self-harm among individuals who interacted with AI-driven treatment programs.
A study from October 2025 at Brown University found that many AI chatbots fall short of meeting essential mental health ethical standards. Additionally, there’s potential for these systems to issue misleading or incorrect diagnoses.
There’s also research suggesting that users might form unhealthy emotional attachments to AI, which can lead to issues like paranoia, emotional instability, and social isolation.
Ginder summed it up by stating, “You can’t put the toothpaste back in the tube. AI is here to stay. The public’s acceptance of these tools is on the rise, especially among younger generations who instinctively search for answers online before asking someone directly.”
“It would be great if regulatory bodies and legal frameworks could catch up with the rapid advancements in AI,” he said.
