Strategic Barriers to Data Sharing in AI Adoption

Dans le cadre des cafés recherche du LIRAES, organisés par Bertrand Chopard, Eric Toulemonde, Université de Namur, viendra présenter un papier intitulé “Strategic Barriers to Data Sharing in AI Adoption“, co-écrit avec Paul Belleflamme et Axel Gautier.

—

Résumé :
We study when competing firms choose to share data with a neutral intermediary that develops an AI-based analytics tool. Data sharing has the nature of a public good, but it also enables costly, firm-specific investments that expand the market and intensify business stealing. We characterize how the interaction between these cooperative and competitive forces–under both strategic complements and strategic substitutes in investment–creates endogenous barriers to data sharing. Our analysis identifies parameter ranges where firms refuse to share data despite clear gains and derives managerial and policy levers to restore efficient data sharing in AI-intensive industries.

—