Gatekeeping in the age of generative AI

Abstract

Twenty years after open access transformed scholarly publishing, another game-changer in the form of generative AI emerged, bringing with it significant implications for the quality of the scholarly record. The peer-review system, which has long been the most important gatekeeping and evaluation mechanism used by publishers, is challenged by reviewer fatigue and shortages and now faces additional pressure from AI-enabled research integrity violations and AI-generated peer reviews. Drawing on data from 47 qualitative interviews conducted with publishers and publishing-adjacent stakeholders, this poster presents insights into how AI has affected editorial quality mechanisms across the industry. In this context, academic publishers are split into two groups: Smaller- and mid-sized presses that mainly rely on peer review for quality control as AI tools are lacking. Large publishing houses that supplement peer review with AI-powered software to help identify and address research integrity issues stemming from AI (ab)use. For publishers, access to AI-detection tools – or the lack thereof – will have long-term effects on the quality of the scholarly record, which may also impact the academic publishing ecosystem in the future. The two-pronged discussion question becomes: what does review-based gatekeeping look like in the age of AI, and how can all publishers, together with the scholarly community, ensure that the quality of their publications is preserved?