Rise of AI-generated open-access review articles after ChatGPT: A case example in inflammatory bowel disease.

Area of Science:

• Biomedical Literature Analysis
• Artificial Intelligence in Science
• Scientific Publishing Ethics

Background:

• Biomedical review articles are crucial for medical education, evidence synthesis, and clinical decisions.
• The emergence of large language models (LLMs) like ChatGPT poses potential risks to the quality and reliability of scientific content due to AI-generated reviews.
• Concerns exist regarding the increasing prevalence and disclosure of AI-authored scientific publications.

Purpose of the Study:

• To investigate the trend of AI-written review articles in inflammatory bowel disease (IBD) following the release of ChatGPT.
• To assess the quality and authorship of potentially AI-generated reviews in the IBD field.
• To highlight the need for safeguards to maintain the integrity of biomedical review articles.

Main Methods:

• PubMed search for open-access IBD reviews published in three periods: pre-ChatGPT, first year post-ChatGPT, and the subsequent year.
• Random selection of 60 articles from each period for analysis.
• Utilized an AI text detector to identify potentially AI-written articles and examined authorship for IBD expertise and disclosure of AI use.

Main Results:

• No AI-written articles were detected in the pre-ChatGPT period.
• AI-written articles constituted 1.7% in the first year post-ChatGPT and 6.7% in the subsequent year.
• None of the identified AI-written articles disclosed AI use, and few had IBD experts as authors.

Conclusions:

• The proportion of AI-written review articles in IBD is increasing, though still relatively low.
• The lack of disclosure and expert authorship in AI-written reviews raises concerns about scientific quality and trustworthiness.
• Urgent establishment of safeguards is necessary to uphold the standards and integrity of biomedical review articles.