• The Agencies make clear that competitors coordinating through the use of a uniform pricing formula engage in per se unlawful price fixing. This follows remarks made by Deputy Attorney General Lisa Monaco at the 39th National Institute on White Collar Crime, hosted by the American Bar Association's Criminal Justice Section on March 7, 2024, regarding the DOJ's enforcement priority of disruptive technology risks.¹ Companies should be on notice that regulators are heavily scrutinizing pricing algorithms.
• According to the Agencies:

1. Plaintiffs do not need to show direct communications between competitors in an algorithmic price-fixing case in order to establish a violation.
2. Even if the algorithm just provides a starting point for pricing, competitor coordination on that still represents a per se unlawful Section 1 violation regardless of whether the ultimate price deviates from that starting point.

• Recent litigation and enforcement actions emphasize how antitrust risk surrounding algorithms and data tools can be mitigated through compliance by identifying the risk associated with AI and pricing algorithms, continuously evaluating how information is exchanged through these tools, conducting regular audits, and creating safeguards to protect commercially sensitive data. 

In depth

In two parallel consolidated class actions against major casino hotel operators in both Las Vegas, Nevada, and Atlantic City, New Jersey, plaintiffs allege that the hotel operators engaged in a price-fixing conspiracy through the use of a pricing algorithm platform offered by one defendant. In both cases, the plaintiffs allege that the algorithm is installed into the defendant-hotels' booking systems and continuously analyzes pricing and occupancy data for each of its subscribers to provide recommendations on guest room pricing, which are updated multiple times per day. Both cases include allegations that defendants utilizing this platform possessed a market share of over 70% in the respective cities. Plaintiffs further allege that the goal of the algorithm was to achieve supra-competitive prices that resulted in overcharges for guests, and, at times, also vacant rooms for the hotels.

Late last year, the US District Court of Nevada dismissed the Las Vegas action without prejudice for three reasons:

1. Plaintiffs failed to sufficiently plead an agreement amongst the defendants to actually use the same pricing algorithm.
2. Plaintiffs' allegation that the defendant-hotels used a defendant's software pricing recommendation 90% of the time was inadequate because the final prices could not be fully derived from the platform recommendations and a 10% rejection rate suggests there was not a complete agreement.
3. Merely identifying the "hotel operators" as participants in the conspiracy meant that the complaint did not sufficiently identify the individual participants involved or the time when the conspiracy began. 

Earlier this year, the defendants in the New Jersey action filed a motion to dismiss, citing heavily to the Las Vegas victory. The defendants argued that the complaint failed to plausibly allege circumstantial evidence of a conspiracy because there are no facts to show that the defendant-hotels engaged in parallel pricing and plaintiffs failed to allege sufficient plus-factors to infer an agreement. Defendants then argued that there was ample evidence to the contrary (i.e., evidence showing defendant-hotels rejected the algorithm's pricing recommendations and participants subscribed to the platform months apart from each other, etc.).

The Agencies' Statement of Interest disagrees with the Las Vegas dismissal, and highlights aspects that the Agencies believe were wrongly decided. The Agencies identified "two legal errors that defendants appear to make in their motion to dismiss".² First, the Agencies argued that plaintiffs do not need to demonstrate direct communications between competitors to plausibly allege an agreement under Section 1 of the Sherman Act.³ The Agencies noted that the Defendant-hotels could have acted in concert by using the same algorithm provider as the "middle man" and "a tacit agreement can be inferred from an invitation proposing collective action followed by a course of conduct showing acceptance".⁴ Second, the Agencies argued that a Section 1 violation can still be established where the Defendants have fixed the starting point of prices, even if the algorithm's pricing recommendations are not binding on the defendants' final rental rates.⁵ The Agencies stated that "both types of agreements [competitors agreeing to fix the final prices and competitors agreeing to fix the starting point for pricing] corrupt the decentralized price-setting mechanism in the market, whether or not they ultimately succeed in raising or stabilizing prices".⁶

In addition to these Statements of Interest, the Agencies have used other forums to comment on AI and algorithmic pricing. For example:

• In June 2017, the FTC, together with the DOJ, submitted a paper to the Organisation for Economic Co-operation and Development (OECD) titled "Algorithms and Collusion".⁷ The paper addressed how to apply and limit US antitrust analysis to business conduct involving technologically advanced tools such as pricing algorithms.
• In June 2022, the FTC issued a report to Congress detailing its concerns with AI tools and signaled that its work in the technological space "will likely deepen as AI's presence continues to rise in commerce".⁸
• In February 2023, the DOJ Principal Deputy Assistant Attorney General, Doha Mekki, called for further scrutiny of pricing algorithms, acknowledging that "[i]n some industries, high-speed, complex algorithms can ingest massive quantities of 'stale,' 'aggregated' data from buyers and sellers to glean insights about the strategies of a competitor" and concluded that "[i]ndeed, we are experiencing an inflection point in the use of algorithms, data at scale, and cloud computing".⁹
• During the American Bar Association 2023 Antitrust Spring Meeting, Leslie Wulff, Chief of DOJ Antitrust Division’s San Francisco office noted that algorithms are not "some sort of, like, set it and forget it, and come back in 50 years and see what the robots have done," adding that "companies have a responsibility to check in on their algorithms, see what they're learning, see how they're adjusting to current market realities".¹⁰
• More recently, on March 7, 2024, Deputy Attorney General Lisa Monaco announced at the 39th National Institute on White Collar Crime hosted by the American Bar Association's Criminal Justice Section that the DOJ's Criminal Division will now "incorporate assessment of disruptive technology risks — including risks associated with AI — into its guidance on Evaluation of Corporate Compliance Programs." She explained that, moving forward, prosecutors will be directed to consider how a company's compliance program assesses and mitigates the risk of misuse of AI.¹¹

Conclusion