AI Pricing: Adoption of Artificial Intelligences and Collusive Price

Abstract

With the growing integration of artificial intelligence (AI) in determining pricing strategies, there is an increasing concern about its potential to foster collusive behavior. Harrington (2012, 2018) underscores the challenge: if AI proves more adept at tacit collusion than humans or if AI-driven collusion is inherently tacit, then it presents a significant hurdle for prosecution under the prevailing interpretation of US antitrust laws. Validating these concerns, Assad et al. (2020) observed collusive price surges linked to the adoption of pricing algorithms among German gas stations. Drawing from game theory—specifically the repeated game paradigm—this paper crafts a foundational mathematical model to analyze competition versus collusion dynamics. It also evaluates the resultant welfare implications of both scenarios. The paper further delves into the broader challenges posed by AI-powered pricing and advocates for potential policy countermeasures, including algorithmic regulation and collusion detection mechanisms.