Attitudes toward artificial intelligence among individuals with anxiety, depression, bipolar disorders, and schizophrenia.

BACKGROUND: Artificial intelligence (AI) increasingly supports medical diagnosis, interventions, and clinical decision-making. In various domains of human communication, AI systems based on large language models offer effective tools for conducting interviews, supportive dialogue, and engaging in problem-focused discussions with healthy individuals and clinical patients. In contemporary clinical settings, interactions between psychiatrists and patients are increasingly complemented by AI; however, the position of this new agent within the therapeutic triad largely depends on patients' attitudes towards AI technologies. OBJECTIVE: The study aimed at two primary objectives: (1) to examine how clinically relevant adaptive and maladaptive personality traits are associated with positive and negative attitudes toward AI, and (2) to evaluate these attitudes among patients diagnosed with anxiety, depression, bipolar disorder, and schizophrenia. METHODS: This multicenter study used the General Attitudes Toward Artificial Intelligence Scale to examine how social adaptation and maladaptation are associated with a broad set of standardized measures, including the DSM-5 Personality Inventory, Schizotypal Traits Questionnaire, Anxiety Sensitivity Index, and Self-Concept Clarity Scale, in psychiatric patients and healthy participants. RESULTS: Patients with schizophrenia and bipolar disorder showed a more positive attitude toward AI use than those with depression and anxiety disorders. A negative attitude towards AI was associated with affective, cognitive, and behavioral vulnerability. This included elevated levels of negative affectivity, detachment, disinhibition, psychoticism, and anxiousness, as well as higher scores on schizotypal traits and reduced self-coherence. CONCLUSION: Patients' symptomatology and diagnostic profiles significantly shape their attitudes toward artificial intelligence, influencing their acceptance or rejection of AI-assisted interventions.