Antitrust Law in the AI Age: Navigating the New Frontier

Introduction

As a trial lawyer and litigator experienced in the complexities of antitrust law, clients increasingly ask me to solve a modern legal conundrum: How does antitrust law--formulated over 100 years ago and designed to regulate human conduct--apply in the era of Artificial Intelligence (AI)? AI is indeed the new frontier. As it becomes ubiquitous, human presidents, antitrust enforcers, judges, juries, and human-led companies are struggling to answer some fundamental legal questions. Will the same antitrust laws that apply to human conduct apply to AI's conduct? Which companies or humans should properly be held liable for anticompetitive conduct by AI? The data so far--including recent litigation involving AI tools for algorithmic pricing as well as President Biden's recent Executive Order calling for regulatory AI standards--suggests that humans and human-led companies may incur antitrust liability when they use or allow AI to do things that would violate the antitrust laws had they been done exclusively by humans. Here, I explore (1) the ongoing policy shift toward regulating and enforcing AI's conduct and (2) companies' liability for allegedly anticompetitive conduct by AI that they use.

Developments 1. Policy Shift Regarding AI.

Congress laid down the foundations of the antitrust laws in the United States in the 1890s and early 1900s, when the government sought to protect free enterprise from the railroad and oil "trusts" of the day who controlled both production and distribution. The antitrust laws, principally the Sherman Act, Clayton Act, and Federal Trade Commission Act, were designed to help promote fair competition. In turn, fair competition drives innovation and brings choice, quality, and market pricing to consumers. At least, that is the idea. Over 100 years later--in a brave new world with technological advances not dreamed of by the Sherman and Clayton Act Congresses--will these same antitrust laws apply to AI's conduct?

The consensus so far is in the affirmative. The rationale is straightforward: If AI's conduct--whether or not guided by human input--results in anticompetitive outcomes, such as price-fixing or unlawful monopolization, it should not escape antitrust scrutiny and potential liability.

The Department of Justice's (DOJ) recent shift in approach exemplifies this perspective. In February 2023, the DOJ Antitrust Division announced a significant change in its review of companies' information sharing practices. This change, which moved away from previous antitrust safe harbors, shows a departure from overly formalistic views of information exchanges. It underscores the concern that tacit coordination facilitated by information sharing could lead to higher prices or suppressed wages. Such a stance reflects a growing acknowledgment of AI's potential role in allegedly anticompetitive information-sharing practices, such as algorithmic pricing.

President Biden's Oct. 30, 2023 Executive Order, calling for regulatory standards for AI, further shapes this landscape. The Executive Order signifies a governmental acknowledgment of the growing influence of AI and the need for a regulatory framework. It suggests an approach where AI is not only subject to existing laws but also potentially to new regulations specifically tailored to its unique characteristics. The Executive Order could lead to a dual regulatory approach: applying existing antitrust laws while also developing new standards and guidelines for AI's use in business practices. Such guidelines might include requirements for transparency in AI algorithms and greater scrutiny of AI's decision-making processes.

2. Recent Litigation Alleging Anticompetitive Conduct by AI.

There is a trend toward seeking to hold human-led companies accountable for AI's alleged anticompetitive conduct. The rationale is that companies cannot evade liability by outsourcing decision-making to AI. Companies using AI tools must ensure that these tools do not engage in conduct that would constitute antitrust violations if performed by humans. Recent litigation, particularly involving AI-driven algorithmic pricing, is illuminating. In separate class action litigation against Yardi and RealPage, property management software providers, putative classes of renters allege that significant corporate landlords used Yardi's and RealPage's respective revenue management software and AI tools to fix rental prices in the multifamily housing market. The allegations are that by sharing non-public, competitively sensitive data, the landlords are colluding to artificially inflate the prices of multifamily residential real estate in the United States. The plaintiffs allege that Yardi and RealPage provided shared software and data analytics to the landlords, who allowed Yardi's and RealPage's AI algorithms to set rental pricing most of the time, thus maximizing rental revenue by restraining competition and harming renters. See, e.g., In re: RealPage Inc. Rental Software Antitrust Litigation (No. II), 3:23-md-03071 (M.D. Tenn. filed April 12, 2023); Duffy v. Yardi Systems Inc., et al., 2:23-cv-01391 (W.D. Wash. filed September 8, 2023).

In the RealPage multidistrict litigation, the DOJ filed a Nov. 15, 2023 Statement of Interest of the United States in which it argues for denial of pending motions to dismiss the action. Specifically, the DOJ argues that the AI's conduct is per se unlawful price-fixing and that the antitrust analysis should be the same whether or not "a software algorithm is involved." Statement, ECF 628, at 15. As courts make rulings in this nascent litigation, antitrust law applicable to algorithmic pricing will take further shape. Perhaps landlords who make some of their own final pricing decisions will not be liable for alleged pricing-fixing by algorithms.

For now, companies should be cautious and recognize that they may be held liable for deploying an AI system that engages in practices arguably violating the antitrust laws. And it remains to be seen whether companies will be held liable when AI acts "independently"--that is, without explicit human inputs or direction--and anticompetitively.

2. Recent Litigation Alleging Anticompetitive Conduct by AI.

There is a trend toward seeking to hold human-led companies accountable for AI's alleged anticompetitive conduct. The rationale is that companies cannot evade liability by outsourcing decision-making to AI. Companies using AI tools must ensure that these tools do not engage in conduct that would constitute antitrust violations if performed by humans. Recent litigation, particularly involving AI-driven algorithmic pricing, is illuminating. In separate class action litigation against Yardi and RealPage, property management software providers, putative classes of renters allege that significant corporate landlords used Yardi's and RealPage's respective revenue management software and AI tools to fix rental prices in the multifamily housing market. The allegations are that by sharing non-public, competitively sensitive data, the landlords are colluding to artificially inflate the prices of multifamily residential real estate in the United States. The plaintiffs allege that Yardi and RealPage provided shared software and data analytics to the landlords, who allowed Yardi's and RealPage's AI algorithms to set rental pricing most of the time, thus maximizing rental revenue by restraining competition and harming renters. See, e.g., In re: RealPage Inc. Rental Software Antitrust Litigation (No. II), 3:23-md-03071 (M.D. Tenn. filed April 12, 2023); Duffy v. Yardi Systems Inc., et al., 2:23-cv-01391 (W.D. Wash. filed September 8, 2023).

In the RealPage multidistrict litigation, the DOJ filed a Nov. 15, 2023 Statement of Interest of the United States in which it argues for denial of pending motions to dismiss the action. Specifically, the DOJ argues that the AI's conduct is per se unlawful price-fixing and that the antitrust analysis should be the same whether or not "a software algorithm is involved." Statement, ECF 628, at 15. As courts make rulings in this nascent litigation, antitrust law applicable to algorithmic pricing will take further shape. Perhaps landlords who make some of their own final pricing decisions will not be liable for alleged pricing-fixing by algorithms.

For now, companies should be cautious and recognize that they may be held liable for deploying an AI system that engages in practices arguably violating the antitrust laws. And it remains to be seen whether companies will be held liable when AI acts "independently"--that is, without explicit human inputs or direction--and anticompetitively.

Conclusion and Best Practices

In the new frontier, where AI becomes ubiquitous, humans and human-led companies may incur antitrust liability for AI's conduct to the extent it mirrors unlawful anticompetitive conduct by humans. Companies employing AI must thus remain vigilant to ensure that AI tools are designed, used, and monitored in a way that complies with the antitrust laws. Regular audits and compliance checks are advisable. Companies should also stay abreast of technological advancements and legal developments, including new regulations and case law that could impact their use of AI. The journey through antitrust compliance and litigation in this new frontier is just beginning--and it promises to be as challenging as it is intriguing!

John S. Gibson is a partner and co-chair of US Business & Commercial Litigation at DLA Piper. Rose Zeck and Quanrico Gibson served as research assistants. CoCounsel served as a generative AI legal assistant.