Commenters debate whether AI-generated writing threatens peer review or exposes weaknesses that were already present. Some argue that submitting AI-written referee reports is disrespectful to researchers and offers editors no value; they want strong sanctions for reviewers and authors who pass off generated text as their own. Others distinguish AI assistance with wording, such as helping non-native English speakers, from AI replacing human judgment, and ask how human contribution should be assessed. The proposed use of AI detectors draws sharp criticism: skeptics say unreliable tools could spur false accusations and witch hunts, while supporters of the paper’s approach see flags as grounds for closer scrutiny. Several commenters also favor refundable deposits or fees to discourage low-quality submissions and compensate reviewers.
Others see unpaid, unevenly motivated reviewing and journal incentives - not AI - as the deeper problem. They describe papers accepted after repeated resubmissions, reviews that miss basic points, and fields where reviewers cannot verify whether results work in practice. One commenter defends peer review as a useful filter, while stressing that it is not designed to detect fraud and cannot catch every flawed study. Views also diverge on what comes next: some expect AI systems to become better reviewers and make human review obsolete, while another favors machine-native output with clear provenance. Calls for clearer, less verbose papers reflect a broader concern that AI can inflate text without improving substance.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.