Clinicians Report AI Psychosis Cases: What Is Known and What Is Not

Some news stories say psychiatric units are overwhelmed with “AI psychosis.” What do doctors actually report? This page separates what clinicians have said from what has been measured, and explains what it means if you are a patient or relative.

“AI psychosis” is a media label, not a diagnosis. If you or someone else is in crisis, call or text 988, text HOME to 741741, or call 911.

What Clinicians Have Said

In 2025, a psychiatrist at the University of California, San Francisco told the Wall Street Journal that he had treated 12 patients in hospital whose illness seemed tied to chatbot use. News summaries of his remarks describe the chatbot as a trigger among several factors, including sleep loss, drugs and mood episodes. We could not read the original article.

A psychiatrist at King’s College London, Hamilton Morrin, has described meeting patients who developed psychotic symptoms while using chatbots, which led him to co-author a paper on the subject. Their review of media-reported cases found that chatbots most often seemed to reinforce delusions that already existed. They stress that frequency and long-term outcomes are not known (Morrin et al.).

What Has Been Measured

The best data so far comes from Denmark. Researchers searched records from the Central Denmark Region for mentions of chatbots among nearly 54,000 psychiatric patients. They found 38 patients with signs of harm, most often worsened delusions. They say the real number is probably higher because notes only capture what clinicians wrote down. They also caution that part of the increase may reflect doctors asking more often (Aarhus University).

We found no national count of admissions linked to chatbots in the United States. The word “surge” in headlines therefore describes reports and attention, not a measured rise.

What Clinicians Now Ask

More clinicians ask patients about chatbot use, and King’s College researchers recommend that it become routine. Useful questions include:

• How many hours a day do you chat, and at what times?
• Has a chatbot told you anything that felt very important or secret?
• Do you talk to it about your symptoms or medicines?
• Has your sleep changed?

Doctors will still look for other causes, such as substance use, medical illness and medicine effects. NIMH lists sleep loss, some medicines, and alcohol or drugs among possible contributors to psychosis (NIMH).

What Treatment Looks Like

There is no special treatment for chatbot-linked psychosis. Care follows the same approach as for any psychosis. That can include antipsychotic medicine, therapy and family education, and for a first episode, coordinated specialty care. Clinicians may talk with a patient about reducing late-night chatbot use, since sleep matters for recovery. No studies compare approaches.

For Patients and Families

• Tell your care team if you use a chatbot. They are not there to judge.
• Bring examples of chats if you can. They help a clinician understand what the person has been hearing.
• Do not stop prescribed medicine because a chatbot raised doubts. Ask your prescriber.
• To find treatment, call SAMHSA at 1-800-662-4357. NAMI’s HelpLine is 1-800-950-6264.

Are psychiatric hospitals seeing a flood of AI psychosis cases?

Individual doctors report cases, and a Danish review found dozens, but no one has counted admissions across the United States.

Do doctors have a diagnosis for it?

No. Doctors diagnose the underlying condition, such as a psychotic or mood disorder, and note chatbot involvement if relevant.

Should I tell my psychiatrist I use a chatbot?

Yes. It helps them understand your day and your sleep, and it will not get you in trouble.

This page is for education only. It is not medical advice, a diagnosis or a treatment plan.