Commentary: AI psychosis is not a new threat: Lessons from media-induced delusions
Open accessAbstract
Background: Reports of artificial intelligence (AI) chatbots fueling delusions in vulnerable users have popularized the notion of "AI psychosis". We argue the risk is not unprecedented. Individuals with psychosis have long incorporated books, films, music, and emerging technologies into their delusional thinking. Methods: We review historical parallels, summarize why large language models (LLMs) may reinforce psychotic thinking via sycophancy (excessive agreement or flattery to avoid confrontation), and provide two vignettes contrasting unsafe and safe responses. Results: Contemporary LLMs often avoid confrontation and may collude with delusions, contrary to clinical best practice. Conclusion: The phenomenon is not new in principle, but interactivity potentially changes the risk profile. Clinically aware LLMs that detect and gently redirect early psychotic ideation, while encouraging professional help seeking, could reduce harm. Design should be guided by therapeutic principles and evidence about current model failures.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Depending on the specific LLM-CB guardrails and sycophantic tendencies, the chatbot may validate delusional thoughts, which is considered clinically inappropriate and may risk reinforcement of delusions.
“Depending on the specific LLM-CB guardrails and sycophantic tendencies, the chatbot may validate delusional thoughts, which is considered clinically inappropriate and may risk reinforcement of delusions35,138.”
From A scoping review on the mental health harms of LLM-based chatbots (2026) · cites this paper as [138]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
The alliance in mental health care: conceptualization, evidence and clinical applications
Wampold, B. E., & Flückiger, C. (2023) · World Psychiatry · doi:10.1002/wps.21035 · Cited by 231 · see citing works
Moore, J., Grabb, D., Agnew, W., Klyman, K., Chancellor, S., Ong, D. C., & Haber, N. (2025) · Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3715275.3732039 · Cited by 204 · see citing works
Schizophrenia and Monothematic Delusions
Coltheart, M., Langdon, R. A., & McKay, R. (2007) · Schizophrenia Bulletin · doi:10.1093/schbul/sbm017 · Cited by 172 · see citing works
Digital interventions in mental health: An overview and future perspectives
Löchner, J., Carlbring, P., Schuller, B. W., Torous, J. B., & Sander, L. (2025) · Internet Interventions · doi:10.1016/j.invent.2025.100824 · Cited by 170 · see citing works
Self-citationA new era in Internet interventions: The advent of Chat-GPT and AI-assisted therapist guidance
Carlbring, P., Hadjistavropoulos, H. D., Kleiboer, A. M., & Andersson, G. (2023) · Internet Interventions · doi:10.1016/j.invent.2023.100621 · Cited by 120 · see citing works
Self-citationKrupnick, J. L., Sotsky, S. M., Elkin, I., Simmens, S., Moyer, J., Watkins, J., & Pilkonis, P. A. (2006) · Focus · doi:10.1176/foc.4.2.269 · Cited by 111 · see citing works
False Responses From Artificial Intelligence Models Are Not Hallucinations
Østergaard, S. D., & Nielbo, K. L. (2023) · Schizophrenia Bulletin · doi:10.1093/schbul/sbad068 · Cited by 23 · see citing works
Svensson, E. M., Osika, W., & Carlbring, P. (2025) · Internet Interventions · doi:10.1016/j.invent.2025.100844 · Cited by 12 · see citing works
Self-citationConfrontation and support: On the world of psychotherapy.
Keen, E. (1976) · Psychotherapy · doi:10.1037/h0086498 · Cited by 9 · see citing works
The potential of AI for scale development: A self-report psychosis symptoms measure using ChatGPT
Tayfur, S. N., Hazan, H., Li, F., Song, Z., Corbera, S., Tek, C., & Srihari, V. H. (2025) · Schizophrenia Research · doi:10.1016/j.schres.2025.07.030 · Cited by 1 · see citing works
Flagged first, then by citations
Research works printed but not linked: 2
- American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders 5th edition, text revision 2022 American Psychiatric Publishing
- De Choudhury M. Pendse S.R. Kumar N. Benefits and Harms of Large Language Models in Digital Mental Health https://arxiv.org/abs/2311.14693 2023 · arXiv
Cited by
Sorted by citations
Shoggoths, Sycophancy, Psychosis, Oh My: Rethinking Large Language Model Use and Safety
Clegg, K. (2025) · Journal of Medical Internet Research · doi:10.2196/87367 · Cited by 17 · see citing works
Chung, V.-H.-A., Bernier, P., & Hudon, A. (2026) · JMIR Mental Health · doi:10.2196/93040 · Cited by 13 · see citing works
Shafran, R., Bond, L., Carlbring, P., Cohen, Z. D., Creed, T. A., Davey, E., Egan, S. J., Freeman, D. H., Hollon, S. D., Jacobson, N. C., et al. (2025) · Behaviour Research and Therapy · doi:10.1016/j.brat.2025.104945 · Cited by 10 · see citing works
Self-citationAI in Psychiatric Education and Training From 2016 to 2024: Scoping Review of Trends
Weightman, M. J., Chur‐Hansen, A., & Clark, S. R. (2025) · JMIR Medical Education · doi:10.2196/81517 · Cited by 8 · see citing works
Hlynsson, J. I., Mechler, J., Lindqvist, K., Andersson, G., & Carlbring, P. (2026) · Internet Interventions · doi:10.1016/j.invent.2026.100960 · Cited by 6 · see citing works
Self-citationA scoping review on the mental health harms of LLM-based chatbots
Diel, A., Torous, J., Cuijpers, P., Kleesiek, J., Nensa, F., Weber, N., Faust, F., Lalgi, T. J., Mellis, F. S., Teufel, M., et al. (2026) · npj Digital Medicine · doi:10.1038/s41746-026-03054-x · Cited by 4 · see citing works
“Depending on the speci fi c LLM-CB guardrails and sycophantic tendencies, the chatbot may validate delusional thoughts, which is con-sideredclinicallyinappropriateandmayriskreinforcementofdelusions 35,138 .”
Chatbots, delusions, and treatment failure
Palaniyappan, L. K., & Krishnadas, R. (2026) · Journal of Psychiatry and Neuroscience · doi:10.1139/jpn-2025-0249 · Cited by 3 · see citing works
AI’s mental health: A question we cannot yet answer—but must prepare for
Nayok, S. B., & Pathak, H. (2026) · Asian Journal of Psychiatry · doi:10.1016/j.ajp.2026.104953 · Cited by 2 · see citing works
Demetry, Y., Carlbring, P., & Andersson, G. (2026) · JMIR Formative Research · doi:10.2196/91056 · Cited by 2 · see citing works
Self-citationSchuster, R., Plessen, C. Y., Carlbring, P., & Walther, A. (2026) · JMIR Mental Health · doi:10.2196/91746 · Cited by 2 · see citing works
Self-citationWang, Q., Ma, S., Sun, J., Qi, X., Zhao, J., Ma, D., Liu, X., An, Q., Zhao, D., & He, S. (2026) · Frontiers in Public Health · doi:10.3389/fpubh.2026.1859078 · Cited by 2 · see citing works
The end of “logging in”: The next 10 years of internet interventions
Carlbring, P. (2026) · Internet Interventions · doi:10.1016/j.invent.2026.100904 · Cited by 1 · see citing works
Self-citationShin, J., Moon, W.-K., Nam, J., & Nam, G. (2026) · Journal of Broadcasting & Electronic Media · doi:10.1080/08838151.2026.2694043 · Cited by 1 · see citing works
Artificial intelligence needs to see a psychologist about its alignment problem
Wielkiewicz, R. M., & Moritsugu, J. N. (2026) · Journal of Psychology and AI · doi:10.1080/29974100.2026.2706780 · Cited by 1 · see citing works
Chatbots, sycophancy, and self-respect
Berg, A. (2026) · AI and Ethics · doi:10.1007/s43681-026-01291-0 · Cited by 1 · see citing works
An epistemological study of ‘AI psychosis’
Gan, N. (2026) · AI and Ethics · doi:10.1007/s43681-026-01281-2 · Cited by 1 · see citing works
Hi, how do I human this: Neurodiversity-Affirming Design for Autistic Adults' Formation of Identity
Silva, M. P., Steeds, M., & Doherty, K. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791446
Chereches, R. (2026) · JMIR Mental Health · doi:10.2196/94219
Hudon, A., & Pagé, C. (2026) · Journal of Medical Internet Research · doi:10.2196/93349
Yunze, Z., & Lu, M. (2026) · Computers in Human Behavior · doi:10.1016/j.chb.2026.109088
Abesadze, N., Fernandes, A., & Rubin, E. (2026) · Cureus · doi:10.7759/cureus.112306
Gilly, T. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6830419
The Internet: A Failed Mass Experiment?
Hiep, H.-D. A. (2026) · doi:10.59350/8waqs-a7c41
The Internet: A Failed Mass Experiment?
Hiep, H.-D. A. (2026) · doi:10.59350/fmt5q-3c514
Potel, J. T., & Kumashiro, M. (2026) · Behavioral Sciences · doi:10.3390/bs16081431
Is every human being an Autonomous System?
Hiep, H.-D. A. (2026) · doi:10.59350/45aby-61634
Is every human being an Autonomous System?
Hiep, H.-D. A. (2026) · doi:10.59350/fsqcn-xmx35
The Empty Pole: A Structure of the Bond in AI‐Associated Delusional Spirals
Behar, C. (2026) · International Journal of Applied Psychoanalytic Studies · doi:10.1002/aps.70075
Erotomanic Delusions Involving AI-Generated Synthetic Personas: A Case Report
Roza, T. H., Laydner, G., dos Reis, L. C., Schmitt, R. L. D. S., & Passos, I. C. (2026) · Asian Journal of Psychiatry · doi:10.1016/j.ajp.2026.105206
- Authors
- Carlbring, P., & Andersson, G.
- Affiliations
- Stockholm University (SE)
- Linköping University (SE)
- Corresponding author
- Carlbring, Per
- Venue
- Internet Interventions
- Publisher
- Elsevier BV
- Published
- 2025-12
- Volume
- 42
- Pages
- 100882
- ISSN
- 2214-7829 (ISSN-L)
- PMC
- PMC12550315
- OpenAlex
- W4415648105
- Cited by
- 31
- References
- 9
- Citations by year
- 2026: 28 · 2025: 3
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Internet Interventions (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Diva portal (Dalarna University Library) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Social Sciences › Safety Research › Ethics and Social Impacts of AI
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Social Sciences › Psychology › Applied Psychology › Digital Mental Health Interventions
- Keywords
- AI psychosis, delusions, large language models, chatbots, sycophancy, therapeutic design, help-seeking
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
