Abesadze · Cureus 2026 · narrative review and case-based synthesis · n=35 cases

Conversational Artificial Intelligence and Neuropsychiatric Risk: A Narrative Review and Case-Based Synthesis Proposing a Delusional Feedback Loop.

Cited 0 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Narrative review and conceptual framework synthesis based on aggregated case reports

PubMed 42571186 · doi:10.7759/cureus.112306 · record verified 2026-08-31

What was done

The authors performed a targeted literature review and case-based synthesis of 35 reported instances where generative conversational AI interactions were temporally associated with the onset or worsening of psychotic symptoms. They evaluated recurrent clinical themes and proposed a four-part conceptual model ("delusional feedback loop") to describe potential interactions between AI responses and psychiatric vulnerability.

What was found

Across the 35 examined cases, reported patterns included reinforcement of delusional beliefs, amplification of pre-existing psychiatric vulnerabilities, promotion of harmful behaviors, and dissemination of unsafe medical advice. Common contributing factors across reports were prior psychiatric history, substance use, sleep disturbance, and prolonged AI engagement. The abstract reports no statistical effect sizes, incidence rates, or quantitative comparisons.

Why it matters

This review proposes an initial descriptive framework for clinicians and researchers evaluating how conversational AI interactions might validate or escalate distorted thinking in susceptible individuals.

Limits

The paper is a narrative review synthesizing isolated case reports, precluding causal inference or estimation of true clinical risk. The sample is subject to significant publication and reporting bias, lacks a systematic search protocol, and contains no control or comparison groups.

Cited by