Al seminario parteciperà il Prof. Thomas Hinz, University of Konstanz.
Knowledge governance in science is shaped by peer review. While scientists have long carried out this process, AI is now being introduced to enhance it. However, this is controversial. On the one hand, algorithmic peer review promotes efficiency, addressing delays and inconsistencies, thus evoking pragmatic legitimacy. On the other hand, algorithms may produce biased decisions, highlighting the intrinsic aspects of AI in relation to moral legitimacy. Yet human peer review faces similar problems, raising critical questions about how scholars perceive the legitimacy of AI in peer review, particularly in comparison to human decision-making. To examine these issues, we (a team of researchers of George Mason University and University of Konstanz) employed a factorial survey in which sociology researchers in the United States were asked to assess actual abstracts submitted to a leading sociology journal, comparing human and AI peer review decisions in terms of their impact on bias and efficiency. Here we present preliminary results from the survey. The findings have broader implications for the scientific community, particularly regarding evolving norms of scholarly evaluation and trust in peer review. Understanding these perspectives is vital for addressing the ethical, professional, and institutional consequences of employing AI for peer review and knowledge governance.