Establish the causal role of chatbots in psychosis-associated presentations

Establish whether chatbot interaction causes or exacerbates psychotic symptoms, rather than merely providing delusional content in individuals whose underlying risk factors drive the observed presentations.

Background

The authors caution that the label “AI psychosis” implies causation that has not been demonstrated. Existing reports are based on anecdotal media coverage, individual clinical case reports, and preliminary observational data, making it difficult to distinguish chatbot-related causation from the influence of pre-existing vulnerability or other risk factors.

The paper specifically notes that the chatbot may function as the content or context of a delusion without being its cause. Longitudinal cohort studies and denominator data from intensive users without apparent harm are identified as necessary to clarify this causal relationship.

References

Another reason to refrain from using the label of AI psychosis'' (orAI-induced psychosis'') is that the term implies a degree of causation which we are, as yet, unable to conclusively ascribe. Given that the current evidence base consists of anecdotal media articles, individual clinical case reports, and preliminary observational data, it is difficult to ascertain the extent to which underlying risk factors for psychotic illness may be driving the observed presentations, with the chatbot serving as delusional content rather than cause.

— An Echo Chamber of One: Should AI Psychosis Be a Distinct Clinical Entity?  (2608.23937 - Yeung et al., 25 Aug 2026) in Section Should AI-Associated Psychosis be a Distinct Clinical Entity?