MinervAI: Using Generative AI to Assist, Not Replace Humans in Peer Review
Published in AI in Science (AIS) 2025 – Copenhagen, Denmark, 2025
In this paper, we argue for a constrained and verification-focused role of LLMs in scientific peer review, emphasizing that evaluative judgments must remain with human experts. Drawing on expert interviews and a review of current challenges, we outline why LLMs are suited only for narrow, objectively checkable tasks. We then present MinervAI, an open-source tool designed to support reviewers by automating citation verification and generating argumentation graphs, thereby alleviating key pain points while preserving human responsibility and oversight.
Recommended citation: Hüsing, I., Petersen, T., Weiher, M.-A., Wolff, C., & Musslick, S. (2025). MinervAI: Using Generative AI to Assist, Not Replace Humans in Peer Review. AI in Science (AIS) 2025.