Survey of 149 people with Parkinson's across 19 countries found 65% had used generative AI (40% for disease-specific questions) with themes of informational, interpretive, and preparational use, and patients bringing AI-generated information into clinical consultations.

Direct therapeutic-discovery value is low, but the findings matter for clinical translation and research conduct because clinicians and trialists must address and validate patient-sourced AI outputs to preserve shared decision-making, data quality, and safety.

Patients' use of generative AI (GenAI) independently from healthcare is increasing across diseases. Little is known about its use among people with Parkinson's disease (PwP). This exploratory convenience-sample, mixed-methods online survey (n = 149, 19 countries) explored PwP's use of GenAI. Among our respondents, 65% had used GenAI, of which 40% had used it for disease-specific inquiries. Qualitative analysis identified informational, interpretive, and preparational uses of GenAI. As PwP increasingly bring AI-assisted data to consultations, clinicians must now actively engage in discussions about these tools, to support shared decision-making, safety and transparency.