Conference leadership index as a novel indicator to measure the scholarly impact of academic institutions in the artificial intelligence field

Traditional institutional evaluation relies heavily on citation data, overlooking the crucial gatekeeper roles within academic conferences. To address this, we propose a novel indicator: the Conference Leadership Index (CLI). The CLI measures scholarly impact beyond traditional bibliometrics by incorporating key structural factors, including conference authority, leadership position level, committee scale, and conference diversity. For the empirical study, we collected data from top artificial intelligence conferences (2012–2021) selected via Research.com impact scores. We compared institutional CLI rankings against publication and patent data. The results demonstrate that CLI effectively reflects scholarly impact but captures a distinct dimension of influence; it correlates moderately with traditional metrics (coefficients of 0.357 with h-index and 0.303 with total citations) and shows almost no correlation with patent performance (coefficient of -0.007), confirming its complementary value. The empirical analysis identifies top AI institutions, revealing a domain dominated by corporate tech giants. Geopolitically, the USA maintains an absolute advantage in conference leadership, even as China thrives and potentially surpasses the USA in Web of Science paper volume. To our knowledge, this is the first comprehensive indicator to integrate conference gatekeeping roles into scientific evaluation. The CLI offers a useful, generalizable tool for diverse fields, shedding light on impact beyond citations and enriching increasingly complex scientific evaluation ecosystems.