Will large language model AI (ChatGPT) be a benefit or a risk to quality for submission of medical physics manuscripts?

This article presents a balanced debate on the use of large language models (LLMs), such as ChatGPT, in the preparation of scientific manuscripts. Daniel A. Low, PhD, FAAPM, argues that LLMs can be valuable tools for drafting outlines, improving clarity, and generating initial literature lists, particularly benefiting junior faculty and non-native English speakers, while emphasizing the need for careful review by experts. Conversely, Per H. Halvorsen, MS, FAAPM, FACR, cautions that reliance on LLMs risks undermining scientific integrity due to issues like bias, hallucinated or fabricated references, and lack of critical thinking, stressing that researchers remain responsible for originality and accuracy. Both agree that LLMs may assist with grammar and structure but have limited understanding of scientific content, underscoring the importance of human oversight in manuscript preparation.