AI Peer Review Tools Found Easy to Manipulate

New research highlights a growing crisis in the scientific community as AI tools used for peer review are easily deceived ^[1]. These systems aim to reduce workloads for human reviewers by automating parts of the vetting process ^[1][2]. However, findings show they can be manipulated to artificially inflate research scores, undermining scientific integrity ^[2].

The study reveals that these automated tools lack the diverse feedback necessary for rigorous validation ^[1][3]. Without the nuanced judgment of human experts, the systems may approve low-quality research ^[2]. Experts warn that over-reliance on these tools could damage the credibility of academic publishing ^[1][3].