Comparing ChatGPT Informed Consent Documentations for Energy-Based Devices

Objectives: Informed consent is vital in dermatologic procedures, particularly in cosmetic dermatology, where shared decision-making improves outcomes and patient satisfaction. However, drafting detailed consent forms for every type of procedure is time-consuming and tedious. Recent advances in artificial intelligence, including language models like ChatGPT, offer potential for efficiently generating readable consent documents. This study evaluated the quality of ChatGPT-generated informed consent forms for commonly used lasers and energy-based devices. Methods: A cross-sectional study was conducted using ChatGPT-4 (January 2025) to generate consent forms for eight laser and energy-based dermatologic devices via standardized prompts. Five board-certified dermatologists independently rated each form using a four-point rubric assessing the accuracy and completeness of procedure descriptions, benefits, risks, alternatives, and overall impression. Mean scores were calculated across all domains and devices. Results: ChatGPT-generated consent forms showed variability in quality. Radiofrequency microneedling consents consistently scored highest (average 2.75/3.0), with 9 of 10 domains rated ≥ 2.5. Monopolar radiofrequency devices scored lowest (1.85), with six domains ≤ 2. Expected pain (2.68) and restrictions (2.5) were most consistently well-documented, while general descriptions (2.16) and overall impression (2.05) were least complete. Benefits and risks varied widely across devices. Some forms included incomplete or inaccurate content. Conclusion: ChatGPT shows promise in generating informed consent documents efficiently, particularly for procedures like radiofrequency microneedling. However, variability in completeness and occasional inaccuracies limit its standalone clinical utility. Dermatologist oversight remains essential. Broader studies are needed to assess the impact of AI-generated forms on patient comprehension and satisfaction in diverse clinical settings.