AI Peer Review in Minutes

The author-level analyses are based on ORCiD identifiers, which are available for only 8.8% of PLOS ONE authors (p. 2). Please quantify how this incomplete disambiguation might bias your results, and consider sensitivity analyses or alternative approaches to address potential distortions in network structure and enrichment calculations.

The study's foundational premise relies on the "suspected paper mill corpus," which is stated to be aggregated from "multiple corpora curated by experts." Please expand the "Suspected Paper Mill Corpus" section in the Methods to briefly summarize the primary criteria or indicators used by the "experts" to curate these corpora.

The manuscript makes strong claims about "coordination" and "concerted attempts at scientific fraud" based on anomalous rates of retractions, PubPeer comments, and short turnaround times. In the discussion, please propose alternative explanations for these anomalous rates (e.g., specialization in high-risk fields, or less diligent but not necessarily fraudulent editorial practices) and how the study's design or subsequent analyses account for or rule out these possibilities.