The peer review process is a mainstay for informing publication decisions at many journals and conferences. It has several strengths that are well-accepted, such as providing a signal about the quality of published papers. Nonetheless, it has several limitations that have been documented extensively, such as reviewer biases affecting paper appraisals. To date, attempts to mitigate these limitations have had limited success. Accordingly, I consider how developments in artificial intelligence technologies—in particular, pretrained large language models with downstream fine-tuning—might be used to automate peer reviews. I discuss several challenges that are likely to arise if these systems are built and deployed and some ways to address these challenges. If the systems are deemed successful, I describe some characteristics of a highly competitive, lucrative marketplace for these systems that is likely to emerge. I discuss some ramifications of such a marketplace for authors, reviewers, editors, conference chairs, conference program committees, publishers, and the peer review process.
"The Other Reviewer: RoboReviewer,"
Journal of the Association for Information Systems, 25(1), 85-97.
Available at: https://aisel.aisnet.org/jais/vol25/iss1/8
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.