Editorial Commentary: Experts in Shoulder Surgery Do Not Consistently Detect Artificial Intelligence–Generated Scientific Abstracts

Abstract

There has been exponential growth in the number of artificial intelligence (AI)– and machine learning (ML)–related publications in recent years. For example, in the field of shoulder and elbow surgery, there was a 6-fold increase in the number of publications between 2018 and 2021. AI shows the potential to improve diagnostic precision, generate precise surgical templates, direct personalized treatment plans, and reduce administrative costs. However, although AI and ML technology has the ability to positively impact biomedical research, it should be closely monitored and used with extreme caution in the realm of research and scientific writing. Current large language models raise concerns regarding the veracity of AI-generated content, copyright and ownership infringement, fabricated references, lack of in-text citations, plagiarism, and questions of authorship. Recent research has shown that even the most experienced surgeons are unable to consistently detect AI-generated scientific writing. Of note, AI detection software is more adept in this role. AI should be used with caution in the development and production of scholarly work.