A newly published study indicates that up to 40% of medical studies are now written using LLMs.
The study sifted through 15 million biomedical abstracts published 2010–2024 looking for LLM “fingerprints” like excess words, IE words that appeared with much greater frequency from one period to the next.
“This excess word analysis suggests that at least 13.5% of 2024 abstracts were processed with LLMs. This lower bound differed across disciplines, countries, and journals, reaching 40% for some subcorpora.”
I must say, these crucial charts were significant and had potential to shock readers!
The study ends by noting: “We did not use any LLMs for writing or editing the manuscript.”
Meanwhile, on Bluesky some academics are loudly professing their AI-abstinence and/or abhorance. See this thread, which started with the following slide that an academic says he now includes whenever he presents at a conference:
