Large language models (LLMs) have diffused rapidly into academic writing since late 2022. Using the complete population of 109,393 research articles published in \textit{PLOS ONE} between 2019 and 2025, we examine population-level structural publication indicators, including full-text manuscript length, authorship team size, reference volume, and cross-linguistic collaboration, before and after 2022. \textit{PLOS ONE}'s multidisciplinary scope and consistent editorial framework allow cross-field comparison under uniform conditions over an extended period. Manuscript length increased substantially, with gains ranging from 14.8\% among African-affiliated authors and 11.7\% among Asian-affiliated authors to 5.3\% among native English-speaking (NES) authors, cutting the word-count gap by 39\%. More strikingly, non-native English-speaking (NNES) authors reduced both authorship team size, from 6.54 to 6.06 authors, or 7.3\%, and collaboration with NES co-authors, from 17.8\% to 12.2\%, or 36\%, while NES authors remained stable in both team size and collaboration rates. Reference counts increased modestly and uniformly across groups. These findings suggest that post-2022 tools may be reshaping not only how science is written, but who writes it together.

翻译：大语言模型自2022年末起迅速渗透至学术写作领域。利用2019年至2025年间发表在《PLOS ONE》上的109,393篇研究论文的完整数据集，我们考察了2022年前后包括全文手稿长度、作者团队规模、参考文献数量和跨语言合作在内的结构性出版指标。《PLOS ONE》的跨学科范畴与一致的编辑框架允许我们在统一条件下对较长时期内的跨领域数据进行对比。结果显示，手稿长度显著增加，增幅从非洲籍作者的14.8%、亚洲籍作者的11.7%到英语母语作者的5.3%不等，使字数差距缩减了39%。更引人注目的是，非英语母语作者不仅将作者团队规模从6.54人缩减至6.06人（降幅7.3%），还将与英语母语合著者的合作率从17.8%降至12.2%（降幅36%），而英语母语作者在团队规模与合作率方面保持稳定。各群体参考文献数量均呈现温和且均匀的增长。这些发现表明，2022年后的工具可能不仅改变了科学写作的方式，还重塑了谁在共同参与科学写作。