Abstract:The dominance of English in global science has long created significant barriers for non-native speakers. The recent emergence of generative artificial intelligence (GenAI) dramatically reduces drafting and revision costs, but, simultaneously, raises a critical question: how is the technology being adopted by the global scientific community, and is it mitigating existing inequities? This study provides first large-scale empirical evidence by analyzing over two million full-text biomedical publications from PubMed Central from 2021 to 2024, estimating the fraction of AI-generated content using a distribution-based framework. We observe a significant post-ChatGPT surge in AI-assisted writing, with adoption growing fastest in contex…
Abstract:The dominance of English in global science has long created significant barriers for non-native speakers. The recent emergence of generative artificial intelligence (GenAI) dramatically reduces drafting and revision costs, but, simultaneously, raises a critical question: how is the technology being adopted by the global scientific community, and is it mitigating existing inequities? This study provides first large-scale empirical evidence by analyzing over two million full-text biomedical publications from PubMed Central from 2021 to 2024, estimating the fraction of AI-generated content using a distribution-based framework. We observe a significant post-ChatGPT surge in AI-assisted writing, with adoption growing fastest in contexts where language barriers are most pronounced: approximately 400% in non-English-speaking countries compared to 183% in English-speaking countries. This adoption is highest among less-established scientists, including those with fewer publications and citations, as well as those in early career stages at lower-ranked institutions. Prior AI research experience also predicted higher adoption. Finally, increased AI usage was associated with a modest increase in productivity, narrowing the publication gap between scientists from English-speaking and non-English-speaking countries with higher levels of AI adoption. These findings provide large-scale evidence that generative AI is being adopted unevenly, reflecting existing structural disparities while also offering a potential opportunity to mitigate long-standing linguistic inequalities.
| Subjects: | Digital Libraries (cs.DL); Computers and Society (cs.CY); Physics and Society (physics.soc-ph) |
| Cite as: | arXiv:2511.15872 [cs.DL] |
| (or arXiv:2511.15872v1 [cs.DL] for this version) | |
| https://doi.org/10.48550/arXiv.2511.15872 arXiv-issued DOI via DataCite |
Submission history
From: Jialin Liu [view email] [v1] Wed, 19 Nov 2025 21:00:18 UTC (1,792 KB)