xRead - Incorporating Artificial Intelligence into Clinical Practice (March 2026)

Carnino et al.

Page 2

Keywords ChatGPT; Artificial intelligence (AI); Manuscript drafting; Ethics

Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Introduction

The realm of healthcare has undergone a remarkable transformation with the advent of Artificial Intelligence (AI), offering unprecedented opportunities to reshape the analysis of medical data, diagnostic processes, and the delivery of patient care [1–3]. With the ability to process vast amounts of complex information rapidly, AI holds the promise of enhancing medical decision-making, predicting diseases, and personalizing treatment plans [4, 5]. Notably, a surge in interest surrounding large language models (LLMs), a sub-type of AI models, has gained traction recently due to their widespread availability and user-friendly interfaces [6]. In recent years, several innovative LLMs and chatbots have been introduced to the public, including the widely discussed ChatGPT by OpenAI in November of 2022. While ChatGPT stands out for its popularity, it is merely a part of a vast and rapidly evolving ecosystem of AI technologies that are transforming the way scientific communication and manuscript drafting are approached. These tools leverage natural language processing (NLP) technology to understand and generate text, offering new avenues for scientific writing and communication [7, 8]. The introduction of advanced LLMs has sparked ethical debates and discussions surrounding integrity, plagiarism, and authorship in higher education [9– 11]. Within the scientific literature, the growing citations of AI tools in PubMed has notably exceeded one thousand since August 2023, underscoring its growing prominence and highlighting the potential for its integration in healthcare [12–14]. Despite the excitement, the integration of various AI-driven tools in scientific discourse has triggered critical ethical discussions. Several science journals have taken a strong stance against its usage, going so far as to prohibit it due to concerns akin to those associated with plagiarism or image manipulation [15]. Furthermore, apprehensions about the quality of papers generated by AI-assistance have been raised, as it occasionally produces inaccurate statements and provides different answers to the same question. This poses a significant challenge, especially as individuals lacking scientific expertise or time to thoroughly draft a manuscript might be empowered to use LLMs for assistance in drafting scientific articles [16, 17]. Our interest in examining AI-generated text, is driven by the need to better understand AI’s influence on the field of Otolaryngology within scholarly publications. This study aims to analyze articles in the high-impact journal JAMA Otolaryngology – Head and Neck Surgery (JAMA – Oto), evaluating the extent of AI-generated content both before and after the release of ChatGPT.

Eur Arch Otorhinolaryngol . Author manuscript; available in PMC 2025 November 01.