Key takeaways
- Feeling helped in the moment is real, but it is not the same as evidence that a tool works as treatment.
- A good rule of thumb is to type only what you would be comfortable having stored somewhere you don’t control.
- You don’t have to choose between a chatbot and a therapist, and you don’t have to hide one from the other.
In this article
It’s late, the house is quiet, and the thing you’ve been carrying all day finally has room to surface. You don’t want to wake anyone. You’re not sure it’s “bad enough” to call a therapist. So you open a chatbot and start typing. If that sounds familiar, you are in a lot of company, and there is nothing embarrassing about it.
People talk to general-purpose chatbots like ChatGPT or Claude, and to AI companion apps, about stress, heartbreak, panic and grief. This post isn’t here to scold you. It’s here to help you use these tools with clear eyes, alongside real care rather than in place of it.
How common is it to turn to AI for mental health support?
More common than many people assume. In a KFF poll of U.S. adults conducted in early 2026, about one in six (16%) said they had sought information or advice about their mental health from AI tools in the past year. Among adults ages 18 to 29, it was about three in ten (28%).
Younger people are doing it too. In a national survey of young people ages 12 to 21, conducted in early 2025 and published in JAMA Network Open, 13.1% said they had used generative AI for mental health advice. Among those ages 18 to 21, it was 22.2%.
Why do people reach for a chatbot first?
A chatbot is there at any hour, with no waitlist, no intake paperwork and little or no cost. You can type what you have never said out loud without watching anyone’s face. For someone who feels stigma about therapy, or who has been trying to find a therapist for months, that matters. The researchers behind the JAMA Network Open survey suggested that young people’s high use likely reflects “the low cost, immediacy, and perceived privacy” of AI advice.
What can a chatbot reasonably help with?
From where I sit as a clinician, the more reasonable uses tend to be modest ones. A chatbot can help you put a feeling into words. It can give you general information about things like anxiety, sleep or grief, which you can then check with a professional. It can help you rehearse a hard conversation, or draft what you want to say to a partner, a parent or a therapist. For some people, naming what is going on is the first step toward reaching out to a person.
The research is more cautious. The American Psychological Association’s 2025 health advisory on AI chatbots and wellness apps notes that general-purpose chatbots have little to no evidence base for mental health care, and it advises against relying on them to deliver psychotherapy or psychological treatment. Feeling helped in the moment is real, but it is not the same as evidence that a tool works as treatment.
Where do chatbots fall short?
A chatbot is not a licensed clinician. The APA advisory spells out the difference: licensed providers are bound by a professional code of ethics, are regulated by state licensing boards and are mandatory reporters of potential harm to self or others. A chatbot carries none of those obligations.
Chatbots can also agree with you too easily. According to the APA, many are “designed to validate and agree with users’ expressed views,” while trained clinicians are taught to both support and challenge you in your best interest. If you are caught in a worry loop or a harsh story about yourself, a tool that keeps agreeing can keep you there.
They can also be confidently wrong. The APA notes that many chatbots are trained on vast, unvetted internet data rather than clinically validated information. And they are not built for emergencies: the advisory describes their ability to safely manage someone in crisis as “limited and unpredictable.”
If you are in crisis or thinking about harming yourself, please reach a person, not a chatbot. Call or text 988 (Suicide & Crisis Lifeline), or call 911 in an emergency.
What happens to what you type?
Therapy records kept by your therapist are generally protected by HIPAA, the federal health privacy law. A chatbot you sign up for on your own is usually a different situation. The Federal Trade Commission explains that HIPAA rules do not apply to health information held by anyone who isn’t a HIPAA covered entity or business associate. What happens to your conversations then depends largely on the company’s own privacy policy and settings.
The APA’s advice is practical: be extremely cautious about sharing sensitive information, avoid entering details that identify you, check the privacy settings, and use options to limit data sharing or request deletion. A good rule of thumb is to type only what you would be comfortable having stored somewhere you don’t control.
How can a chatbot and therapy work together?
You don’t have to choose between a chatbot and a therapist, and you don’t have to hide one from the other. The APA advisory recommends that people tell their health providers which AI tools they use, and that clinicians ask about it. That conversation helps a therapist notice when advice you received may be unhelpful or out of step with your care.
In practice, you can bring what you’ve been talking about into session. What felt helpful? What felt off? A therapist can also offer what a chatbot can’t: someone who gets to know you over time, notices your tone and what goes unsaid, and works with you on the patterns underneath the questions.
That is different from a therapist choosing to use AI as a tool inside your care, openly and with your consent, which we wrote about in What Happens When a Therapist Brings AI Into the Room?
When is it worth talking to a person?
Consider reaching out if you keep returning to a chatbot night after night with the same worry, if it has become the main place you go instead of the people in your life, or if anxiety or low mood keeps showing up. None of that means you did something wrong.
Therapy Aligned offers sessions in person in Metuchen, NJ and telehealth across New Jersey. Clients must be located in New Jersey at the time of the session. If worry is what keeps sending you back to the chat window, anxiety therapy in New Jersey describes how we work with it, and you can use our intake form to request a free consultation.
Sources
- American Psychological Association. (2025, November). Use of generative AI chatbots and wellness applications for mental health: An APA health advisory. https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-chatbots-wellness-apps
- Federal Trade Commission. (n.d.). Mobile health app interactive tool. https://search.ftc.gov/business-guidance/resources/mobile-health-apps-interactive-tool
- McBain, R. K., Bozick, R., Diliberti, M., Zhang, L. A., Zhang, F., Burnett, A., Kofner, A., Rader, B., Breslau, J., Stein, B. D., Mehrotra, A., Uscher-Pines, L., Cantor, J., & Yu, H. (2025). Use of generative AI for mental health advice among US adolescents and young adults. JAMA Network Open, 8(11), Article e2542281. https://doi.org/10.1001/jamanetworkopen.2025.42281
- Montero, A., Montalvo, J., III, Kearney, A., Valdes, I., Kirzinger, A., & Hamel, L. (2026, March 25). KFF tracking poll on health information and trust: Use of AI for health information and advice. KFF. https://www.kff.org/public-opinion/kff-tracking-poll-on-health-information-and-trust-use-of-ai-for-health-information-and-advice/




