{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T03:12:53Z","timestamp":1789009973684,"version":"build-2803163510"},"reference-count":72,"publisher":"National Academy of Sciences","issue":"22","license":[{"start":{"date-parts":[[2026,5,29]],"date-time":"2026-05-29T00:00:00Z","timestamp":1780012800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2026,6,2]]},"abstract":"<jats:p>Large language models (LLMs) are rapidly changing academic research, raising questions of who is adopting these tools and under what conditions. This article analyzes full texts of 7.3 million journal articles published from 2020\u20132025 by four major publishers (Elsevier, Frontiers, MDPI, and PLoS) to track the prevalence of LLM-associated language and identify social and institutional correlates of adoption. A corpus of 228 focal words exhibiting sharp post-2022 frequency increases consistent with LLM output was developed; articles were scored on their rate of focal word usage. By 2025, an estimated 57% of published articles exhibited evidence of LLM influence, up from 12% in 2023. Among articles exhibiting LLM-influenced text, there is substantial heterogeneity, ranging from subtle linguistic influence to articles mostly or entirely LLM-generated. Difference-in-differences models reveal that LLM-associated language varies markedly across regions, institutional ranks, publishers, disciplines, and journal tiers. Economic development and proximity to English as a primary language are key predictors of regional variation. Lower-ranked institutions exhibit higher rates than elite universities, young for-profit publishers show elevated rates vis-\u00e0-vis competitors, and academic fields differ widely in adoption. LLM adoption in academic writing is pervasive but socially stratified. As models grow more powerful and their use becomes further entrenched in academic research, understanding social dynamics of adoption will be essential for governing the evolving relationship between AI and academic knowledge production.<\/jats:p>","DOI":"10.1073\/pnas.2605754123","type":"journal-article","created":{"date-parts":[[2026,5,29]],"date-time":"2026-05-29T17:39:58Z","timestamp":1780076398000},"update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":8,"title":["The diffusion of large language models in published academic articles"],"prefix":"10.1073","volume":"123","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0427-9283","authenticated-orcid":false,"given":"Kyle","family":"Siler","sequence":"first","affiliation":[{"id":[{"id":"https:\/\/ror.org\/03dbr7087","id-type":"ROR","asserted-by":"publisher"}],"name":"Department of Leadership, Higher and Adult Education, Ontario Institute for Studies in Education, Data, Equity and Policy in Education Lab, University of Toronto","place":["Toronto"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"341","published-online":{"date-parts":[[2026,5,29]]},"reference":[{"key":"e_1_3_4_1_2","unstructured":"A. Gray ChatGPT \u201ccontamination\u201d: Estimating the prevalence of LLMs in the scholarly literature (2024) https:\/\/arxiv.org\/abs\/2403.16887."},{"key":"e_1_3_4_2_2","unstructured":"W. Liang Monitoring AI-modified content at scale: A case study on the impact of ChatGPT on AI conference peer reviews (2024) https:\/\/arxiv.org\/abs\/2403.07183."},{"key":"e_1_3_4_3_2","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-025-00212-1"},{"key":"e_1_3_4_4_2","doi-asserted-by":"crossref","unstructured":"A. Azaria T. Mitchell The internal state of an LLM Knows When It\u2019s Lying (2023) https:\/\/arxiv.org\/abs\/2304.13734.","DOI":"10.18653\/v1\/2023.findings-emnlp.68"},{"key":"e_1_3_4_5_2","doi-asserted-by":"publisher","DOI":"10.1146\/annurev.soc.24.1.265"},{"key":"e_1_3_4_6_2","volume-title":"Diffusion of Innovations","author":"Rogers E. M.","year":"2003","unstructured":"E. M. Rogers, Diffusion of Innovations (Simon and Schuster, New York, ed. 5, 2003).","edition":"5"},{"key":"e_1_3_4_7_2","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.adt3813"},{"key":"e_1_3_4_8_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41562-025-02273-8"},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1016\/S0048-7333(02)00007-0"},{"key":"e_1_3_4_10_2","doi-asserted-by":"publisher","DOI":"10.1093\/pnasnexus\/pgaf399"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.1093\/pnasnexus\/pgag007"},{"key":"e_1_3_4_12_2","doi-asserted-by":"publisher","DOI":"10.1128\/mBio.00640-12"},{"key":"e_1_3_4_13_2","unstructured":"Retraction Watch The retraction watch leaderboard (2026) https:\/\/retractionwatch.com\/the-retraction-watch-leaderboard\/."},{"key":"e_1_3_4_14_2","unstructured":"T. S. Juzek Z. B. Ward Why does ChatGPT \u201cDelve\u201d so much? Exploring the sources of lexical overrepresentation in large language models (2025) https:\/\/arxiv.org\/abs\/2412.11385."},{"key":"e_1_3_4_15_2","doi-asserted-by":"crossref","unstructured":"V. Rao A. Kumar H. Lakkaraju N. B. Shah Detecting LLM-written peer reviews (2025) https:\/\/arxiv.org\/abs\/2503.15772.","DOI":"10.1371\/journal.pone.0331871"},{"key":"e_1_3_4_16_2","doi-asserted-by":"publisher","DOI":"10.1093\/oso\/9780199240531.001.0001"},{"key":"e_1_3_4_17_2","volume-title":"Atlas of Science","author":"B\u00f6rner K.","year":"2010","unstructured":"K. B\u00f6rner, Atlas of Science (MIT Press, Cambridge, 2010)."},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1002\/asi.24339"},{"key":"e_1_3_4_19_2","volume-title":"Engines of Anxiety: Academic Rankings, Reputation, and Accountability","author":"Espeland W. N.","year":"2016","unstructured":"W. N. Espeland, M. Sauder, Engines of Anxiety: Academic Rankings, Reputation, and Accountability (Russell Sage, New York, 2016)."},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11192-007-2036-x"},{"key":"e_1_3_4_21_2","doi-asserted-by":"publisher","DOI":"10.15626\/MP.2024.4601"},{"key":"e_1_3_4_22_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.respol.2022.104608"},{"key":"e_1_3_4_23_2","unstructured":"D. S. Chawla What\u2019s wrong with the journal impact factor in 5 graphs (2018) https:\/\/www.nature.com\/nature-index\/news\/whats-wrong-with-the-jif-in-five-graphs."},{"key":"e_1_3_4_24_2","unstructured":"Genderize.io Genderize.io (2026). https:\/\/genderize.io\/. Accessed 26 January 2026."},{"key":"e_1_3_4_25_2","unstructured":"H. Yakura Empirical evidence of Large Language Model\u2019s influence on human spoken communication (2024) https:\/\/arxiv.org\/abs\/2409.01754."},{"key":"e_1_3_4_26_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.respol.2018.03.006"},{"key":"e_1_3_4_27_2","doi-asserted-by":"publisher","DOI":"10.1162\/qss_a_00368"},{"key":"e_1_3_4_28_2","doi-asserted-by":"publisher","DOI":"10.7208\/chicago\/9780226000329.001.0001"},{"key":"e_1_3_4_29_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pbio.3002184"},{"key":"e_1_3_4_30_2","doi-asserted-by":"publisher","DOI":"10.2307\/3588284"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.1177\/1056492619835314"},{"key":"e_1_3_4_32_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.adw3000"},{"key":"e_1_3_4_33_2","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-025-01463-8"},{"key":"e_1_3_4_34_2","unstructured":"J. Lee A. J. Conrad Borchers T. J. Alvero R. F. Kizilcec The digital divide in generative AI: Evidence from large language model use in college admissions essays (2026) https:\/\/arxiv.org\/abs\/2602.17791."},{"key":"e_1_3_4_35_2","doi-asserted-by":"publisher","DOI":"10.1287\/orsc.2026.ed.v37.n3"},{"key":"e_1_3_4_36_2","doi-asserted-by":"publisher","DOI":"10.1162\/qss_a_00327"},{"key":"e_1_3_4_37_2","first-page":"222","article-title":"Journals and publishers crack down on research from open health data sets","volume":"390","author":"O\u2019Grady C.","year":"2025","unstructured":"C. O\u2019Grady, Journals and publishers crack down on research from open health data sets. Science 390, 222\u2013223 (2025).","journal-title":"Science"},{"key":"e_1_3_4_38_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pbio.3003152"},{"key":"e_1_3_4_39_2","doi-asserted-by":"crossref","unstructured":"E. Gibney How AI slop is causing a crisis in computer science (2026) https:\/\/www.nature.com\/articles\/d41586-025-03967-9.","DOI":"10.1038\/d41586-025-03967-9"},{"key":"e_1_3_4_40_2","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1145\/1506409.1506410","article-title":"Conferences vs. journals in computing research","volume":"52","author":"Vardi M. Y.","year":"2009","unstructured":"M. Y. Vardi, Conferences vs. journals in computing research. Commun. ACM 52, 5 (2009).","journal-title":"Commun. ACM"},{"key":"e_1_3_4_41_2","doi-asserted-by":"publisher","DOI":"10.1002\/asi.23349"},{"key":"e_1_3_4_42_2","unstructured":"L. Carbone Advancing mathematics research with large language models (2025) https:\/\/www.arxiv.org\/pdf\/2511.07420v1."},{"key":"e_1_3_4_43_2","doi-asserted-by":"publisher","DOI":"10.1093\/pnasnexus\/pgae591"},{"key":"e_1_3_4_44_2","volume-title":"Autonomous Technology","author":"Winner L.","year":"1977","unstructured":"L. Winner, Autonomous Technology (MIT Press, Cambridge, 1977)."},{"key":"e_1_3_4_45_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.adh2586"},{"key":"e_1_3_4_46_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2401227121"},{"key":"e_1_3_4_47_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.adz9311"},{"key":"e_1_3_4_48_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-025-09922-y"},{"key":"e_1_3_4_49_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-024-07146-0"},{"key":"e_1_3_4_50_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.chbah.2025.100207"},{"key":"e_1_3_4_51_2","unstructured":"M. Abdulhai How LLMs distort our written language (2026) https:\/\/arxiv.org\/abs\/2603.18161."},{"key":"e_1_3_4_52_2","doi-asserted-by":"crossref","unstructured":"Z. Sourati A. S. Ziabari M. Dehghani The homogenizing effect of large language models on human expression and thought. arXiv [Preprint] (2026) https:\/\/doi.org\/10.48550\/arXiv.2508.01491 (Accessed 26 January 2026).","DOI":"10.1016\/j.tics.2026.01.003"},{"key":"e_1_3_4_53_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsos.241776"},{"key":"e_1_3_4_54_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2518443122"},{"key":"e_1_3_4_55_2","unstructured":"P. Song P. Han N. Goodman Large language model reasoning failures (2026) https:\/\/www.arxiv.org\/abs\/2602.06176."},{"key":"e_1_3_4_56_2","unstructured":"M. Sharma Towards understanding sycophancy in language models (2025) https:\/\/arxiv.org\/abs\/2310.13548."},{"key":"e_1_3_4_57_2","unstructured":"J. Y. Bo M. Kazemitabaar M. Deng M. Inzlicht A. Anderson Invisible saboteurs: Sycophantic LLMs mislead novices in problem-solving tasks (2026) https:\/\/arxiv.org\/abs\/2510.03667."},{"key":"e_1_3_4_58_2","doi-asserted-by":"crossref","unstructured":"S. Rathje . Sycophantic AI increases attitude extremity and overconfidence (2026). https:\/\/osf.io\/preprints\/psyarxiv\/vmyek_v3.","DOI":"10.31234\/osf.io\/vmyek_v3"},{"key":"e_1_3_4_59_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.aec8352"},{"key":"e_1_3_4_60_2","unstructured":"R. M. Batista T. L. Griffiths A rational analysis of the effects of sycophantic AI (2026) https:\/\/arxiv.org\/pdf\/2602.14270."},{"key":"e_1_3_4_61_2","doi-asserted-by":"crossref","unstructured":"M. Atari M. J. Xue P. S. Park D. E. Blasi J. Henrich Which Humans? (2024) https:\/\/osf.io\/preprints\/psyarxiv\/5b26t_v1.","DOI":"10.31234\/osf.io\/5b26t"},{"key":"e_1_3_4_62_2","doi-asserted-by":"publisher","DOI":"10.1177\/29768624251408919"},{"key":"e_1_3_4_63_2","unstructured":"M. Naddaf AI linked to explosion of low-quality biomedical research papers (2025) https:\/\/www.nature.com\/articles\/d41586-025-01592-0."},{"key":"e_1_3_4_64_2","doi-asserted-by":"crossref","unstructured":"N. Jones ArXiv preprint server clamps down on AI slop (2026) https:\/\/www.science.org\/content\/article\/arxiv-preprint-server-clamps-down-ai-slop.","DOI":"10.1126\/science.aef8896"},{"key":"e_1_3_4_65_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2518075122"},{"key":"e_1_3_4_66_2","doi-asserted-by":"publisher","DOI":"10.1145\/3571730"},{"key":"e_1_3_4_67_2","unstructured":"A. T. Kalai O. Nachum S. S. Vempala E. Zhang Why language models hallucinate (2025) https:\/\/arxiv.org\/pdf\/2509.04664."},{"key":"e_1_3_4_68_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-024-07421-0"},{"key":"e_1_3_4_69_2","unstructured":"Y. Sakai H. Kamigaito T. Watanabe HalluCitation matters: Revealing the impact of hallucinated references with 300 hallucinated papers in ACL conferences (2026) https:\/\/arxiv.org\/abs\/2601.18724."},{"key":"e_1_3_4_70_2","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-026-00969-z"},{"key":"e_1_3_4_71_2","unstructured":"G. Cabanac C. Labb\u00e9 A. Magazinov Tortured phrases: A dubious writing style emerging in science. Evidence of critical issues affecting established journals (2021) https:\/\/arxiv.org\/abs\/2107.06751."},{"key":"e_1_3_4_72_2","doi-asserted-by":"crossref","unstructured":"K. Siler Replication data and code for: The diffusion of large language models in published academic articles. Open Science Framework. https:\/\/osf.io\/btrnj\/. Deposited 17 February 2026.","DOI":"10.1073\/pnas.2605754123"}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.2605754123","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,9]],"date-time":"2026-06-09T15:44:53Z","timestamp":1781019893000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/10.1073\/pnas.2605754123"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,29]]},"references-count":72,"journal-issue":{"issue":"22","published-print":{"date-parts":[[2026,6,2]]}},"alternative-id":["10.1073\/pnas.2605754123"],"URL":"https:\/\/doi.org\/10.1073\/pnas.2605754123","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,29]]},"assertion":[{"value":"2026-02-19","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-04-27","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-05-29","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e2605754123"}}