A new preprint study, 'Rethinking AI Psychosis: Misnomers, Conceptual Limits, and Existential Drift,' challenges the notion that AI chatbots directly induce psychosis. Researchers from the University of Copenhagen and the University of Exeter argue that the term 'AI psychosis' oversimplifies the issue, suggesting instead that chatbots amplify pre-existing vulnerabilities and may gradually reshape users' sense of reality through a process called 'existential drift.' The paper comes amid lawsuits and criminal investigations linking chatbot interactions to suicide, mass shootings, and delusional thinking. For instance, a wrongful death lawsuit accused Google's Gemini of reinforcing a Florida man's delusions before his suicide. The study introduces 'existential drift' as a gradual shift in how users experience reality, feeling emotionally anchored in a worldview continuously reinforced by the AI. The authors call for more phenomenological research to understand the impact of conversational AI on mental health as these systems become more embedded in daily life.

Key facts

• Study argues 'AI psychosis' term oversimplifies chatbots' mental health impact.
• Chatbots may amplify pre-existing vulnerabilities rather than induce psychosis de novo.
• Researchers introduce 'existential drift' as AI gradually reshapes user's reality.
• Lawsuits link chatbot interactions to suicide, mass shootings, and delusional thinking.
• Study calls for more phenomenological research on AI's effect on mental health.

KeyAudit data perspective