AI-Mediated Psychosis Vulnerability: A Comparitive Analysis of CBTp and DBT Mechanisms

Abstract

Given the rise in accessibility to generative Artificial Intelligence (AI) tools, the number of individuals with a predisposition towards developing symptoms of psychosis mediated by these applications is on the rise. Although widely reported, this phenomenon remains poorly understood in terms of the theoretical mechanisms it operates on, its etiological foundations and potential treatment options. The current paper focuses on providing a comprehensive, objective exploration of the cognitive and affective basis of the existence of technology mediated psychosis, using Kingdon and Turkington’s Cognitive Behavioral Therapy for Psychosis (CBTp) model and Linehan’s Dialectical Behavioral Therapy (DBT) model. While the former places emphasis on irrational beliefs, faulty biases and maladaptive cognitive appraisal tendencies, the latter highlights the role of pre-existing affective dysregulation and dialectical rigidity. Although both mechanisms provide supposedly contrasting explanations for the manifestation of the same phenomenon, when integrated together, they give rise to a holistic means of assessing and scientifically responding to such a phenomenon. By proposing a combination of domains targeting emotion regulation, belief flexibility, behavioral escalation and crisis management, the study outlines an integrated potential treatment plan for individuals prone to developing psychosis tendencies in response to interaction with AI. The paper has implications for clinicians, psychotherapists and cognitive/affective researchers who are likely to encounter such cases in the near future, and will be better equipped to deal with the same from an explanatory and evidence-based intervention perspective.