Policies Permitting LLM Use for Polishing Peer Reviews Are Currently Not Enforceable

TL;DR

This study evaluates the enforceability of policies banning LLM use in peer reviews, revealing that current detectors often misclassify human-AI collaborations, thus questioning the reliability of AI detection in academic peer review contexts.

Contribution

The paper systematically assesses the effectiveness of state-of-the-art AI detectors on simulated peer reviews, highlighting their limitations and the challenges in reliably identifying AI-assisted reviews.

Findings

All detectors misclassify some human-AI collaborative reviews as AI-generated.

Incorporating peer-review-specific signals offers limited detection improvements.

Current detectors are unreliable for accurately identifying AI use in peer reviews.

Abstract

A number of scientific conferences and journals have recently enacted policies that prohibit LLM usage by peer reviewers, except for polishing, paraphrasing, and grammar correction of otherwise human-written reviews. But, are these policies enforceable? To answer this question, we assemble a dataset of peer reviews simulating multiple levels of human-AI collaboration, and evaluate five state-of-the-art detectors, including two commercial systems. Our analysis shows that all detectors misclassify a non-trivial fraction of LLM-polished reviews as AI-generated, thereby risking false accusations of academic misconduct. We further investigate whether peer-review-specific signals, including access to the paper manuscript and the constrained domain of scientific writing, can be leveraged to improve detection. While incorporating such signals yields measurable gains in some settings, we identify limitations in each approach and find that none meets the accuracy standards required for identifying AI use in peer reviews. Importantly, our results suggest that recent public estimates of AI use in peer reviews through the use of AI-text detectors should be interpreted with caution, as current detectors misclassify mixed reviews (collaborative human-AI outputs) as fully AI generated, potentially overstating the extent of policy violations.