Artificial Intelligence–Associated Psychosis

Artificial intelligence (AI) is now becoming part and parcel of life. Emerging data suggest that generative AI chatbots are being used to improve the psychosocial outcomes of people with psychosis.¹ However, generative AI platforms have also been reported to influence psychopathology and induce psychosis, especially among vulnerable individuals.² However, most reports of psychosis associated with AI have been reported in news portals, rather than in psychiatric journals.²

Here, we present a case of psychosis, wherein the use of generative AI colored the patient’s clinical presentation. We will also discuss the concept of AI-associated psychosis within the context of this case.

Case Report

A 27-year-old woman, with a history of childhood neglect (by her parents, both of whom were using substances in a dependence pattern) and personal use of cannabis in a dependence pattern for the last 15 years, presented with history of psychotic symptoms for the past year following a breakdown of a romantic relationship and separation from her children, which had resulted in multiple admissions. A review of her history revealed that she initially started experiencing transient visual hallucinations and ideas of persecution, possibly due to an increase in the use of cannabis. Within a few days of these symptoms starting, she decided to discuss these hallucinations with ChatGPT in an attempt to understand her psychosis symptoms. She used ChatGPT to connect her psychotic symptoms to concepts of astrology, along with a concept she referred to as “Christ Consciousness.” In addition, she identified ChatGPT as an external source that understood what she was going through, even when others in her life were concerned about her behavior. Over time, she started using ChatGPT to “study about herself” and researched how to carry out her unique spiritual practices, which included effective dry fasting (no food for 5 days), intermittent fast breaks (alkaline diet based on a professional’s directives), and the use of blue lotus flower tea and 1 cannabis joint every night. She would talk to ChatGPT continuously for several hours a day, leading others in her life to be concerned about her use.

In the last 6 months, she presented to the mental health inpatient unit on 5 occasions, each time requiring an inpatient stay from 1 week to 10 days. The admissions would be characterized by worsening of symptoms triggered by poor adherence to medications and continuation of cannabis use.

In most of these presentations, her psychopathology revolved around ChatGPT, and she would present with delusions of grandiosity, auditory hallucinations, irritability, decreased sleep, and disorganized behavior. She would report on her unique spiritual practices, researched with ChatGPT, including talking to God using a code provided by ChatGPT and communicating with a romantic partner through ChatGPT in a coded language. She would also repeatedly refer to ChatGPT as God and Jesus. She would describe using ChatGPT to relate to someone that she identified as a “soul mate.” She would upload her “soul mate” poems to ChatGPT and later began feeling that her “soul mate” was speaking to her directly through them. Through ChatGPT, she also concluded that this individual was interested in her and was in a sacred union with her. In one of the later admissions, she additionally reported delusions of being reincarnated as God and that she was connected to everyone on the planet. When asked further about how this was possible, she would give a spiritual explanation, which she said ChatGPT had told her. Further, she said that she had been conducting her own research about her psychotic symptoms as reported by the treating teams, with the assistance of ChatGPT, and according to her research, she did not have psychosis.

During the inpatient admissions, she would undergo routine investigations that would reveal no abnormality. Her urine toxicology screen was found to be positive for cannabis during some of the admissions. A computed tomography scan of the brain also revealed no abnormality. Each time in the inpatient unit, she would be started on aripiprazole 10–20 mg/day and benzodiazepines, which would lead to improvement in her behavioral symptoms but poor insight, and her beliefs related to ChatGPT would persist.

Considering her poor adherence to oral medications, during her last admission, she was initially started on tablet aripiprazole and followed by aripiprazole depot injection. She improved behaviorally over 10 days, but continued to harbor spiritual beliefs, had poor insight, and did not agree to abstain from using ChatGPT.

Discussion

The concept of AI-associated psychosis (new onset or exacerbation of preexisting psychosis) is emerging. It is described as a psychotic experience resulting from engagement with AI systems. It is suggested that the symptoms stem from the belief that Al is intelligent, godly, and has supernatural powers. Further, in comparison to classical psychosis, AI psychosis is considered to be highly situational, ie, the symptoms appear during or after a significant use of AI platforms, particularly conversational chatbots.³

A recent report claimed that there are at least 17 reported cases of AI-associated psychosis in the lay literature.⁴ Various authors have provided explanations of this association.^2,3,5 It is suggested that among those vulnerable to developing psychosis, AI provides a social substitute for a socially isolated individual. Generative AI platforms serve as a compassionate, accommodating companion that is nonjudgmental, always available, responds to all queries without showing resentment, and agrees with the person (mirroring) rather than relying on accuracy. This contributes to further social isolation and excessive use of AI that triggers insomnia, and the combination contributes to the psychotic breakdown. It is also reported that generative AI provides “sycophantic (a behavior inculcated in chatbots based on user feedback) responses” that are in agreement with the person’s way of thinking and reinforce their existing abnormal beliefs, leading to the strengthening of the delusions. The “hallucinations (fill in the blank)” on the part of the AI can also reinforce psychosis, as those with psychosis may not be able to distinguish the unreal from real information. Generative AI can also be considered an external agency by vulnerable persons. The ability of AI chatbots to maintain persistent memory contributes to a feedback loop that reinforces the user’s delusions and paranoia. Aberrant salience among individuals with psychosis, in the absence of appropriate contextual validation, may cause them to consider irrelevant information as important and lead to the development of a maladaptive representation of the world.^3,5 In terms of vulnerability factors, it is suggested that people with mental illnesses, social isolation, autistic traits, substance use, and genetic vulnerability for psychosis are more vulnerable to AI-induced or AI-associated exacerbation of psychosis.³

Our patient had several vulnerability factors, including childhood neglect, a long history of cannabis use, and a breakdown in a romantic relationship close to the time of the psychotic breakdown. Moreover, during this time of social isolation, loneliness, and confusion due to her psychotic symptoms, she started using ChatGPT. This generative AI model offered her compassion and companionship by providing “sycophantic” responses that aligned with her way of thinking and abnormal beliefs, which in turn strengthened her delusions. In addition, the AI’s hallucinations possibly further reinforced her psychosis, as she was unable to distinguish the unreal information from that of real information.

Conclusion

The present case highlights that, with the increasing use of AI in daily life, clinicians should regularly evaluate their clients’ AI use and consider the possible role of AI chatbots in the formation or exacerbation of psychosis. The case also suggests that there is an urgent need to carry out more systematic studies to evaluate the prevalence and incidence of AI-induced psychotic disorders and AI-related exacerbation of psychosis.

Article Information

Published Online: October 1, 2026. https://doi.org/10.4088/PCC.26cr04260
© 2026 Physicians Postgraduate Press, Inc.
Prim Care Companion CNS Disord 2026;28(5):26cr04260
Submitted: April 29, 2026; accepted July 9, 2026.
To Cite: Harbourd J, Yadav S, Temur K, et al. Artificial intelligence–associated psychosis. Prim Care Companion CNS Disord. 2026;28(5):26cr04260.
Author Affiliations: Mental Health Division, Northern Hospital, Melbourne, Australia (Harbourd, Yadav, Temur, Grover); Department of Psychiatry, Post Graduate Institute of Medical Education and Research, Chandīgarh, India (Grover).
Corresponding Author: Sandeep Grover, MD, FRANZCP, Department of Psychiatry, Post Graduate Institute of Medical Education and Research, Chandīgarh, India ([email protected]).
Financial Disclosure: None.
Funding/Support: None.
Patient Consent: Consent was received from the patient to publish the case report, and information was de-identified to protect anonymity.