Management Summary Research Study “An Interest-Group Theory of AI-Tool Governance in Science”
This study examines why academic institutions, including scientific journals, funding agencies, university departments, and professional associations, adopt rules that restrict the use of artificial intelligence tools in research. Drawing on a formal interest-group model in the public choice tradition, specifically a Tullock lobbying contest, the study shows that such rules can be understood as the equilibrium outcome of organized conflict within the scientific community. The starting point is the observation that scientists differ in how productively they can complement artificial intelligence tools, and that academic rewards, including publication opportunities in selective outlets, grants, promotions, prizes, and professional recognition, depend largely on relative standing. Even when artificial intelligence tools raise aggregate scientific output, scientists with smaller productivity gains can rationally support restrictions, because restrictions compress the relative advantage enjoyed by colleagues with higher AI complementarity. The model characterizes the equilibrium probability of restriction, lobbying expenditures, and the dissipation of real resources through political contestation. A key implication follows: because rank-based rewards are privately valuable and partly positional, the chosen rules can be excessively restrictive relative to the social optimum. The mechanism applies most strongly where governance bodies retain substantial discretion over rule-setting and where scientists cannot easily avoid restrictive policies by shifting to alternative venues. The study also derives a range of empirically testable predictions that help to distinguish an interest-group account of artificial intelligence governance from a public-interest account. The practical implication is that restrictions on the use of artificial intelligence tools in research deserve critical scrutiny, since the analysis suggests that such restrictions may reflect the organized interests of particular constituencies within the scientific community.
Target groups of stakeholders: Academic administrators, journal editors, officials at funding agencies, science policy makers, professional associations, and researchers across disciplines.
Citation: Berggren, N. (forthcoming). An interest-group theory of AI-tool governance in science. Public Choice.