A U.K.-based academic publisher has introduced what it calls the first artificial-intelligence system designed to spot copied or template-like peer reviews, a growing integrity risk in scholarly publishing. Institute of Physics Publishing said its Duplicate Review Checker screened roughly 500,000 reviewer reports from 2020 to 2025 and flagged nearly 2,500 with at least 60% overlap, including 785 with 80% overlap and 89 exact duplicates. The tool combines semantic matching with character-level comparison to surface suspicious similarities and is now being rolled out across all IOPP journals. The move follows mounting evidence of “review mills” and manipulation: PLOS said it examined about 150 papers for peer-review integrity concerns between 2024 and 2026, retracting 40 and flagging 55. Publishing executives and research-integrity scholars say the problem is likely far larger than known cases, as pressure to publish and the rise of AI-enabled workflows strain traditional safeguards.
Related articles:
