Physician Opinions on Artificial Intelligence Chatbots In Dermatology: A National Online Cross-Sectional Survey of Dermatologists

November 2024 | Volume 23 | Issue 11 | 972 | Copyright © November 2024


Published online July 30, 2024

doi:10.36849/JDD.8239

Julian Cortes BS, Taraneh Paravar MD, Reid Oldenburg MD PhD

Department of Dermatology, University of California, San Diego Health, San Diego

Abstract
Background: Artificial intelligence chatbots (AIC) have sharply risen in popularity. Dermatology, heavily involving visual, clinical, and pathological pattern-recognition techniques, will be impacted by AIC. Thus, this study aims to categorize the attitudes and beliefs of American dermatologists towards AIC and their potential uses, benefits, and risks.
Methods: An online cross-sectional survey was distributed to dermatologists across the United States. Questions explored opinions on AIC along with perceived benefits, risks, and important considerations for the incorporation of AIC into the practice of dermatology. Demographic data and self-reported understanding of AIC were also collected.
Results: 192 complete responses were received. 53.6% of respondents were female. 44.3% were between ages 30 to 39. 41.1% had 0 to 10 years of experience as attending physicians. 76.5% of participants believed it is somewhat or very likely that AIC will be formally incorporated into dermatology. Higher self-reported understanding of AIC was associated with an increased perceived likelihood of AIC implementation as well as decreased perceived risk associated with AIC. Notably, 86% of respondents believed AIC would impact "patient education," while concerns regarding "misinformation" and "incorrect diagnoses" were prevalent (89% and 78.5%, respectively). Participants anticipated AIC's role primarily in administrative tasks, with 75.7% citing "reduced work burden on physicians" as a potential benefit.
Conclusion: Dermatologists in the United States foresee the integration of AIC into their practice, emphasizing its potential in administrative roles. Concerns revolve around the complexity of medical understanding and effective patient communication. As AIC continues to evolve, ongoing studies are crucial to evaluate their safety and efficacy in dermatological practice.

J Drugs Dermatol. 2024;23(11):972-978. doi:10.36849/JDD.8239

INTRODUCTION

New artificial intelligence (AI) tools are being incorporated into the practice of medicine. Research has focused on AI's potential usage in disease diagnosis and patient care, treatment efficacy, and processing of large data.1 One AI tool recently experiencing a sharp rise in use is the AI chatbot (AIC). The release of Chat GPT (Generative Pre-trained Transformer) by OpenAI in November 2022 was received with the fastest growing consumer base for an internet application, reaching 100 million users in two months.2 Incorporating advanced language-learning technology, ChatGPT offers anyone with an internet connection the opportunity to receive a targeted response to virtually any prompt, including those related to medicine.

Dermatology, heavily involving visual, clinical, and pathological pattern-recognition techniques will likely be impacted by AI. Currently, some AI tools have been developed to facilitate certain aspects of dermatological practice such as tracking nevus morphology and evolution, identifying concerning skin lesions, and even producing a differential diagnosis for a provided image.3 Given the broad accessibility of AIC and their potential to process images, these tools may become highly relevant in the future of dermatology.  However, with this possibility comes new concerns: how will AIC impact physicians, patients, or medical care delivery? To better understand the current role and future of AIC in dermatology, our study aimed to identify current opinions held by dermatologists. We deployed a survey designed to categorize the attitudes and beliefs of dermatologists towards AIC along with its potential uses, benefits, and risks for providers and patients.

MATERIALS AND METHODS

Study Population
This study employed an online cross-sectional survey distributed to dermatologists across the United States. Contacted groups included the Association of Professors of Dermatology, the American College of Mohs Surgery, the San