Generative AI at the crossroads of values: competing orders of worth in university and publisher regulations

Abstract

This study employs the orders of worth theoretical framework, developed by Boltanski and Thévenot, to analyze how prestigious universities and academic publishers justify and critique the use of generative artificial intelligence (AI). Through qualitative analysis of AI guidelines from Ivy League universities, top UK universities, and leading publishers, we identify how these organizations mobilize different orders of worth. Civic worth dominates across both sectors, with its emphasis on integrity and accountability. Universities more readily incorporate industrial worth (efficiency) and project/network worth (adaptation) than publishers. Universities establish faculty authorization tests that preserve domestic worth through traditional hierarchies, while creating differentiated legitimacy spaces based on educational context. Publishers construct strict human accountability tests that reject AI as an author and establish compromises that permit limited industrial efficiency for technical functions, simultaneously maintaining rigid boundaries around core scholarly contributions. The tension between industrial efficiency and inspired authenticity creates the central dynamic that institutions must negotiate. Thus, institutional responses to AI are not just technical implementations but social arenas where multiple value systems are negotiated, reinforcing certain conceptions of the common good while marginalizing others.