Finding the Ghost in the Shell:EU and US Antitrust Enforcement of AI Collusion

Abstract

In recent years, considerable focus has been directed toward AI collusion and whether the proliferation of AI-driven decision-making could enable entities to coordinate in anti-competitive ways beyond the scope of enforcement. While the perceived risk may seem exaggerated - antitrust law is well equipped to deal with these questions, and AI may not lead to unavoidable collusive outcomes - the question deserves careful consideration. Notably, US enforcers face challenges in contrast to their EU colleagues, as US case law has developed a narrow notion of “agreement” or “understanding” under high and inconsistent evidentiary burdens. Despite the odds, AI-assisted collusion is currently being challenged more aggressively through ex-post action in the US and discussed in the EU. The outcomes and effects of this new wave of litigation remain to be seen. This paper explores antitrust enforcement of AI collusion to identify potential enforcement gaps and available remedies on both sides of the Atlantic. 

Keywords: AI, collusion, algorithms, big data, Article 101, Section 1