Dermatologists' and Dermatopathologists' Views on Artificial Intelligence in the Clinical Setting: A Systematic Review
Poster #: 178
Session/Time: A
Author:
Bailey Paige Sullivan, BS
Mentor:
Alice Roberts, MD, PhD
Research Type: Review Article
Abstract
INTRODUCTION:
Artificial intelligence (AI) emerged in the second half of the twentieth century, but a significant milestone in 2018 was the introduction of GPT-1, the precursor to ChatGPT. AI has transitioned from theoretical to practical applications in a clinical setting, including analyzing images, drafting visit notes, educating patients, and personalizing treatment plans. However, barriers, such as providers' education and acceptance of new AI developments, hinder the implementation of AI innovations. Dermatologists' views on AI are not well documented. We aim to analyze clinicians' perceptions of AI based on current awareness of its use in the clinical setting and opinions on its role in patient-provider relationships.
MAIN BODY:
Following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines (PRISMA), publications were systematically reviewed from PubMed, Web of Science, Academic Search Complete, Science Direct, CINAHL, and Google Scholar. The search strategy included keywords pertaining to AI, dermatologists, perceptions, and surveys. Two independent reviewers (MW, BS) using Rayyan software blindly conducted the title and abstract review but were not blinded during the full text review. The studies were evaluated according to our inclusion criteria. Two independent reviewers performed the data extraction with minor assistance from ChatGPT. Data analysis was performed using Excel. The risk of bias assessment will be performed by two independent reviewers using ROBINS-I criteria. 4,370 articles were generated in the systematic search. The 32 included studies were published between 2020-2025 and consisted of 32 observational studies, including 24 cross-sectional survey studies. Included studies represent participants from 21 countries, not including one survey study which collected data from 92 countries. Though data interpretation and analysis are currently ongoing, the consensus that AI will have a positive impact in the dermatological field in the next decade appears to exist among most dermatologists.
CONCLUSION:
Preliminary findings from this ongoing systematic review suggest that the majority of dermatologists and dermatopathologists generally view the potential role of AI in dermatology positively. However, despite this optimism, the widespread implementation of AI in clinical settings remains limited. The literature currently provides little insight into the disconnect between these favorable perceptions and real-world adoption. This gap presents a significant opportunity for future research to investigate the underlying barriers, such as workflow integration, costs, ethical considerations, and trust in AI outputs that may hinder the translation of positive sentiment into practice. Addressing these challenges will be crucial for ensuring that AI progresses from a promising adjunctive tool to a widely adopted element of dermatological care.
Artificial intelligence (AI) emerged in the second half of the twentieth century, but a significant milestone in 2018 was the introduction of GPT-1, the precursor to ChatGPT. AI has transitioned from theoretical to practical applications in a clinical setting, including analyzing images, drafting visit notes, educating patients, and personalizing treatment plans. However, barriers, such as providers' education and acceptance of new AI developments, hinder the implementation of AI innovations. Dermatologists' views on AI are not well documented. We aim to analyze clinicians' perceptions of AI based on current awareness of its use in the clinical setting and opinions on its role in patient-provider relationships.
MAIN BODY:
Following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines (PRISMA), publications were systematically reviewed from PubMed, Web of Science, Academic Search Complete, Science Direct, CINAHL, and Google Scholar. The search strategy included keywords pertaining to AI, dermatologists, perceptions, and surveys. Two independent reviewers (MW, BS) using Rayyan software blindly conducted the title and abstract review but were not blinded during the full text review. The studies were evaluated according to our inclusion criteria. Two independent reviewers performed the data extraction with minor assistance from ChatGPT. Data analysis was performed using Excel. The risk of bias assessment will be performed by two independent reviewers using ROBINS-I criteria. 4,370 articles were generated in the systematic search. The 32 included studies were published between 2020-2025 and consisted of 32 observational studies, including 24 cross-sectional survey studies. Included studies represent participants from 21 countries, not including one survey study which collected data from 92 countries. Though data interpretation and analysis are currently ongoing, the consensus that AI will have a positive impact in the dermatological field in the next decade appears to exist among most dermatologists.
CONCLUSION:
Preliminary findings from this ongoing systematic review suggest that the majority of dermatologists and dermatopathologists generally view the potential role of AI in dermatology positively. However, despite this optimism, the widespread implementation of AI in clinical settings remains limited. The literature currently provides little insight into the disconnect between these favorable perceptions and real-world adoption. This gap presents a significant opportunity for future research to investigate the underlying barriers, such as workflow integration, costs, ethical considerations, and trust in AI outputs that may hinder the translation of positive sentiment into practice. Addressing these challenges will be crucial for ensuring that AI progresses from a promising adjunctive tool to a widely adopted element of dermatological care.