How AI Is Changing the Way People Think About Mental Health Support

A lot of emotional pain stays private for many people, and the loneliness around it can make speaking to another person feel harder than saying nothing at all. But AI has moved into daily life faster than expected, and some people are now typing personal concerns into a chatbot that they have never brought to a friend or a therapist. 

John Torous, director of the Digital Psychiatry Division at Beth Israel Deaconess Medical Center, said during a National Academy of Medicine discussion that one in three people have used an AI chatbot for emotional support. 

That same behavior appears in BetterHelp’s 2026 State of Stigma Report, which found that 15% of Americans use AI for mental health support, while 74% said society still discourages people from seeking help. As more people turn to AI for support, clinicians are increasingly considering where these tools may fit alongside care from licensed professionals. 

Why some people are turning to AI first

There are several reasons people turn to AI first when emotional concerns become hard to sort through, and accessibility is one of the easiest to understand. 

Dr. Leana Wen, an emergency physician and clinical associate professor at George Washington University, told CNN that chatbots are available at any hour and respond right away, letting someone type out what they are feeling before a scheduled appointment is possible. 

Wen also points to privacy, noting that some people feel more comfortable typing about a problem than discussing it with a parent or counselor. 

For someone who is not ready to open up, spending money on care creates another hurdle, especially if therapy requires paying out of pocket or dealing with insurance. Arthur C. Evans Jr., chief executive of the American Psychological Association, describes general chatbots as “readily available and easy to access without insurance.”

BetterHelp’s 2026 State of Stigma Report found that 52% of 2,000 U.S. adults named it as a barrier to therapy while 27% pointed to stigma or fear of judgment. Those barriers may help explain why some people experiment with AI before seeking professional care, but whether they continue using it depends on whether the tool proves useful once the conversation begins. 

Where AI can add value in mental wellness

Perhaps what is most fascinating about AI is how useful it becomes around the smaller tasks that often surround mental health care. 

A person heading into therapy may know they have been anxious or overwhelmed but still struggle to explain what has been happening, and Penn State World Campus has described AI as a tool for compiling questions or concerns before meeting with a licensed professional. 

The same need to organize thoughts continues even between appointments, especially when something difficult happens days before a person is scheduled to speak with a therapist. AI-guided journaling gives people a place to write down what happened and how they reacted while the details and feelings are still fresh.

Dr. Mona Barman, director of care transformation at BetterHelp, says day-to-day concerns do not wait for therapy sessions. In that context, responsibly used AI may help people organize their thoughts, reflect on experiences, or document concerns between sessions rather than replace licensed care. Experts have found that writing those experiences down can also make changes in mood easier to follow over time. 

A review of AI applications in mental health found that AI-based apps can monitor mood fluctuations and help identify possible triggers, giving people and clinicians more specific information about when distress appears or what tends to happen around it.

Research has also moved into clinical studies, where teams are testing whether carefully designed AI systems can support people who are already living with diagnosed mental health conditions.

In a clinical trial involving 106 adults, Dartmouth researchers found significant symptom improvements among participants using a purpose-built generative AI mental health chatbot, including a 51% average reduction among participants with depression and 31% among those with generalized anxiety. The findings came from a single study of this specific system and should not be generalized to consumer AI chatbots.

But even with those results, study author Michael Heinse said no generative AI agent is ready to operate fully on its own in mental health, leaving a lot of important questions about safety and reliability unresolved. 

The Limitations AI Cannot Overcome

The biggest issue with AI is that it’s still in its infancy, especially when it comes to mental health. A chatbot can sound caring, but it does not build the same relationship as a therapist who learns how someone communicates over time.

APA CEO Arthur C. Evans Jr. has said general chatbots are “supportive to a fault” and lack the same ability to recognize warning signs as human professionals. And missing one of those signs is especially serious during a crisis, where the response depends on recognizing risk and knowing when immediate human help is needed.

General-purpose AI chatbots should not be relied on for crisis intervention or emergency mental health support; people in immediate danger or experiencing a mental health emergency should seek immediate human help through emergency services or an appropriate crisis resource. 

If you or someone you know is struggling or in crisis, help is available. Text or call 988 or chat 988lifeline.org. Support is available 24/7.

But even before a situation reaches that level, a chatbot struggles to recognize when reassurance is the wrong response. C. Vaile Wright, senior director of the Office of Healthcare Innovation at the APA, has said chatbots simply validate what someone is already doing, even when harmful thinking needs to be challenged. 

Ioana Literat, an associate professor of technology, has warned that people can mistake fluent AI responses for credibility, even when the information itself is wrong, so a user may not realize the response needs to be checked against a qualified source.

Still, public trust leans heavily toward human judgment, with BetterHelp’s 2026 State of Stigma Report finding 74% of people would never trust AI more than a licensed professional. 

Ethical questions will shape the future

Given how complex AI is, sharing personal mental health details with a chatbot raises immediate questions about where that information goes and who may have access to it. APA guidance says people receiving psychological services should be told when AI is being used, and tools handling sensitive behavioral health data need strong privacy protections.

Knowing an AI system is involved, however, does not settle who is responsible if its guidance causes harm. APA says psychologists remain responsible for final decisions, while legal responsibility is less settled when a general chatbot operates outside clinical care. 

Trusting AI with mental health support also depends on whether the tool has been tested well enough to work fairly across different groups. APA recommends rigorous validation before AI tools are used in psychological practice, especially when incomplete data may produce biased results.

The need to prove that a tool works fairly, while protecting the personal data it relies on, is also leading regulators to consider new rules around transparency and data use. But even with new rules under consideration, many people still want a person involved. BetterHelp’s 2026 State of Stigma Report found 46% said AI-based tools work best paired with a medical professional. 

The future may be human care enhanced by AI

Struggling with mental health is a deeply human experience, and the tools people use to get through it are changing faster than the field can study them. George Nitzburg, an assistant professor of clinical psychology at Teachers College, Columbia University, has said technology reduces barriers so people can get something rather than nothing, with the goal being to widen access and help people find support before things reach a crisis.

And younger adults tend to see AI the same way, with 42% of Millennials calling it a stepping stone or supplement to human care. One place that supporting role is especially clear is in the work clinicians already do.