WASHINGTON, DC – October 30: US President Joe Biden hands the pen to Vice President Kamala Harris … [+]
Curbing misconduct resulting from the misuse of artificial intelligence is now a top federal priority. Anticompetitive conduct, subject to antitrust laws, is one of the potential AI-related harms highlighted in the October 2023 Presidential Executive Order on AI. Antitrust laws apply to review a variety of business practices, including competitor pricing of goods and services, mergers, and conduct by “dominant” companies, including technology platforms. As commercial applications of AI rapidly expand, business leaders and executives need to know what enforcers are looking out for to avoid pitfalls that could expose them and their companies to significant legal liability.
In his State of the Union address last week, President Joe Biden criticized the misuse of algorithms to manipulate prices, saying:[f]”Or for millions of renters, we are cracking down on large landlords who violate antitrust laws by fixing prices and inflating rents.” Algorithmic anticompetitive behavior is already the subject of extensive civil litigation and growing government investigations, but regulators need to tread carefully to ensure that scrutiny does not stifle economically beneficial business applications of algorithms.
AI Fundamentals and Competitive Risks
According to Oracle content strategist Michael Chen, an AI model is “both the set of selected algorithms and the data used to train those algorithms to make the most accurate predictions.” The rapid growth in computing power in recent years has made AI an ever-more powerful tool used in business planning and management.
However, executives should keep in mind that providing information to train AI software could raise legal issues, including antitrust issues as well as copyright infringement, data privacy and security violations, and anti-discrimination law violations. In-house counsel responsible for monitoring legal risk exposure should work closely with company management to avoid such issues.
The order recognizes that while AI has the potential to bring great benefits, it also poses great risks: When competition is harmed, the solution is antitrust enforcement, and the order states, “addresses the risks arising from the centralization of control of key inputs.” [and] “We will take steps to prevent unlawful collusion and prevent dominant companies from disadvantaged their competitors.”
Antitrust Enforcement and AI
The federal antitrust enforcement agencies, the Department of Justice and the FTC, are responsible for addressing three specific competition-related concerns mentioned in the AI Order: conspiracies such as price fixing, concentration of control of key inputs, and dominant firms that disadvantage competitors. While the Department of Justice challenges collusive agreements as antitrust crimes (the FTC has no criminal enforcement authority), the two agencies share enforcement authority to prosecute civil anticompetitive agreements, monopolies, and anticompetitive mergers. Private parties can also seek treble damages for anticompetitive arrangements.
Antitrust claims over AI-related activities are still in their infancy. Currently, the Department of Justice and the FTC are focusing on algorithm-related collusion, and a recent study by antitrust scholar Satya Marar highlights the potential and pitfalls of antitrust claims in this area.
AI Algorithm Conspiracy Case
According to the U.S. Supreme Court, collusive cartel conduct – secret agreements between business rivals to fix prices – is “antitrust law's greatest evil.”
Without an agreement, there would be no criminal antitrust violation. But enforcement authorities are concerned that AI algorithms could carry out such collusion without a specific agreement. AI algorithms that are “trained” on industry pricing practices could help companies predict how competitors are likely to set prices in the future. If companies buy the same algorithmic software, they could avoid price wars and immediately respond to price changes made by competitors, helping to stabilize and lock in prices.
Two civil antitrust cases involving alleged collusion through algorithms have attracted the government's attention.
In Real Page In the currently pending litigation, renters allege that multiple landlords independently fed nonpublic business information into the same pricing algorithm created by RealPage and used it to set rental prices. “The alleged scheme meets the legal standard for unlawful price fixing per se,” the DOJ argued in a November 2023 court filing. This explicit statement suggests that the DOJ is seeking criminal charges raising similar fact patterns.
The Department of Justice and the FTC filed a joint amicus brief in November 2023. McKenna Duffy v. Yardi Systems, Inc. The lawsuit alleges that the plaintiff tenants unlawfully agreed to their landlords “using Yardi's pricing algorithms to artificially inflate rental prices for their multi-family properties.”
In a March 2024 article on the FTC's “Business Blog,” staffers Hannah Garden Monheit and Ken Marber emphasized that “agreements to share pricing recommendations, lists, calculations, or algorithms may still be illegal even if the conspirators retain pricing discretion or otherwise violate the agreement.” The clear message is that enforcement agencies will be aggressive in investigating algorithmic pricing.
Algorithmic pricing raises tough questions, but could be beneficial
The conditions under which parallel conduct, including the selection and training of AI algorithms, can provide the basis for a criminal conviction raise new and challenging litigation questions. Algorithms may seem to make it easier to understand and maintain price-related decisions. At the same time, as Professors Joshua Davis and Anupama Reddy point out in a report published by the University of San Francisco, algorithms may create a record of the procedures, inputs, and calculations underlying price decisions, making it easier for the Department of Justice to identify and prosecute conspiratorial conduct.
There are positive aspects to deploying AI algorithms in business. As Marar points out (citing academic research), algorithms can better customize product offerings to reflect consumer preferences and enable companies to respond to pricing changes more quickly and efficiently. More efficient pricing allows companies to respond more effectively to consumer demand.
Regulators should take these advantages into account when developing enforcement policies to deter misuse without discouraging the use of economically desirable algorithms that stimulate competitiveness.
