Which scientists are using AI-generated language in their research papers?

Embargoed until: Publicly released: 2025-08-04 19:00

International

Photo by Aerps.com on Unsplash

The use of large language models in the writing of academic papers has risen sharply over recent years, with computer scientists and researchers from countries with fewer English-language speakers more likely to use it, according to international research. The team analysed over 1 million preprints and published papers across arXiv, bioRxiv and Nature portfolio journals from 2020 to 2024 looking for likely AI usage, and found its use likely increased by 22% for computer science papers while mathematics papers and the Nature portfolio (the publisher of this paper) saw smaller increases up to 9%. The researchers say a pattern of more frequent AI usage in countries with fewer English speakers could be a symptom of the prevalence of English in scientific research and the challenges this poses for these countries.

Journal/
conference: Nature Human Behaviour

Research:Paper

Organisation/s: Stanford University, USA

Funder: We thank D. A. McFarland, D. Jurafsky, Y. Yin, Z. Izzo, X. V. Lin, L. Chen and H. Ye for their helpful comments and discussions. J.Z. is supported by the National Science Foundation (grant nos. CCF 1763191 and CAREER 1942926), the US National Institutes of Health (grant nos. P30AG059307 and U01MH098953) and grants from the Silicon Valley Foundation and the Chan-Zuckerberg Initiative. H.L. is supported by the National Science Foundation (grant nos. 2244804 and 2022435) and the Stanford Institute for Human-Centered Artificial Intelligence (HAI).

Media Contact/s

Contact details are only visible to registered journalists.