University-Industry Knowledge Transfer: The Case of the Godfathers of Artificial Intelligence

Abstract

Though university-industry linkages are channels for two-way knowledge sharing, in the case of breakthrough research, it is the path from scientific discovery to commercial application that is of primary interest. The central question is whether knowledge can be codified and disseminated in disembodied form, or if engagement with knowledge creators, either explicit engagement through collaborative relationships or potential engagement made possible by geographic proximity, is required. We consider the case of the godfathers of artificial intelligence (AI), Geoffrey Hinton (University of Toronto), Yoshua Bengio (Université of Montréal) and Yann LeCun (New York University), who collectively won the 2018 Turing Award, the Nobel prize of computing, and examine the factors that affect the ability of private technology developments to absorb and build upon their research. We construct a dataset consisting of the modern AI patent universe and use patent citations to godfather research as an indicator of knowledge transfer, modelling the likelihood that an AI patent will cite godfather research as a function of geographic proximity and social relations. We find strong effects for geographic proximity, where inventors were former co-inventors or co-authors of the godfathers, and for absorptive capacity as a crucial control.