Study examines how large language models are changing peer review in AI academic publishing

A recent study explores the impact of large language models on the peer review process in academic publishing, particularly in AI conference proceedings. It investigates whether LLMs are reshaping the essential evaluative roles of peer review, which traditionally seeks to enhance manuscript quality through clarity and originality assessments. Researchers scrutinized shifts in peer review reports since the rise of LLMs, concentrating on detailed changes in linguistic features such as word and sentence length and complexity. Additionally, the study looks into how LLMs influence the linguistic structure, evaluative emphasis, and recommendation signals in review comments. While earlier studies indicate that LLMs are starting to affect peer review, this research thoroughly assesses the magnitude of these transformations. The rapid evolution of LLMs has led to significant disruptions in academic communication, particularly in the evaluation and refinement of scholarly work prior to publication. This paper, referenced as arXiv:2604.19578v1, serves as a cross-disciplinary examination of these changing dynamics in academic evaluation systems.