Unlawful Unilateral Use of AI: When Algorithms Become Instruments of Anticompetitive Conduct

In this piece for the StepAhead Antitrust blog, authors Robert Klotz, Lee Berger, Weisiyu Jiang, Rachel Carlo & Domniki Mari (Steptoe Antitrust) dive into how unilateral uses of AI can raise antitrust issues across the EU and US. With competition authorities increasingly focused on how digital platforms leverage algorithms, the authors examine the potential risks around self-preferencing, personalized pricing, and predatory conduct. They highlight that both jurisdictions are moving beyond coordinated conduct and algorithmic collusion theories to scrutinize independent, AI-driven strategies that may distort market outcomes.

The discussion begins with self-preferencing, noting that dominant platforms’ control over data and ranking systems can enable them to favor their own products or those of preferred partners. In the EU, Article 102 TFEU and the Digital Markets Act provide clear frameworks to address such conduct, with landmark cases against Google and Amazon underscoring growing regulatory scrutiny. Similarly, in the US, the DOJ’s recent success in challenging Google’s search practices shows enforcers are prepared to apply existing law to AI-enabled ranking and default-setting behavior.

The article then turns to algorithmic price discrimination, where advanced tools allow firms to segment consumers and personalize pricing in ways that could extract maximum willingness to pay. While this may increase efficiency, it can also raise exploitative abuse concerns when deployed by dominant firms. The authors note emerging enforcement attention on pricing algorithms, citing the FTC’s challenge to Amazon’s Project Nessie as a key example of how US regulators may rely on Section 5 of the FTC Act to address algorithmic strategies that fall outside traditional collusion frameworks.

Finally, the authors explore AI-driven predatory pricing, emphasizing that algorithms can dynamically implement below-cost strategies in targeted ways that are harder to detect and evaluate under existing legal tests. In the EU, this could test the limits of the AKZO standard, while in the US, such tactics may revive interest in predatory pricing theories and spark renewed attention to the Robinson-Patman Act. The piece concludes that as AI-enabled unilateral conduct evolves, regulators are increasingly prepared to adapt longstanding competition laws to new digital-market realities, raising future compliance and enforcement challenges for firms deploying advanced pricing and ranking tools…

CONTINUE READING…