Applications of generative artificial intelligence in chronic pain: a scoping review.

Abstract

Chronic pain is a complex biopsychosocial condition that imposes substantial clinical and socioeconomic burdens worldwide. Although digital health interventions increasingly support chronic pain management, the role of generative artificial intelligence (AI) remains insufficiently characterized. This scoping review mapped applications of generative AI in chronic pain and identified research gaps. PubMed, Scopus, and Web of Science were searched on October 19, 2025, using "chronic pain" combined with five named generative AI systems. Primary research studies published in English that focused on generative AI applications in chronic pain were included. Of 402 records identified, 25 studies met the eligibility criteria. The studies were classified into five areas: patient and family education and self-management support; evaluation of medical information quality, accuracy, and reliability; pain assessment, analysis, and monitoring; clinical and educational support for healthcare professionals; and interpretation of patient-drawn images. Studies demonstrated technical feasibility across several applications, but most evidence was based on model performance, expert ratings, or retrospective data rather than prospective clinical outcomes. Limitations included variable accuracy, readability, AI hallucinations, inconsistent reporting, and limited external validation. Therefore, generative AI may ultimately serve as an adjunct to clinical judgment, but its clinical benefit and safety remain to be established.