This research introduces a novel approach to detect when peer reviews are generated by LLMs instead of human experts, preserving the integrity of scientific validation processes.

• Created a first-of-its-kind benchmark dataset of human and AI-generated peer reviews
• Developed new detection methods specifically calibrated for the peer review domain
• Achieved high accuracy in distinguishing between human and LLM-written reviews
• Revealed concerning vulnerabilities in current peer review systems against LLM exploitation

For the security community, this work addresses emerging threats to scientific integrity by providing tools to identify AI-generated content masquerading as expert human evaluation, helping maintain trust in research validation mechanisms.

Original Paper: Is Your Paper Being Reviewed by an LLM? A New Benchmark Dataset and Approach for Detecting AI Text in Peer Review