AI Risks Hindering DOJ Antitrust Cases With Complex Pricing Data

At a Dec. 13 Judiciary Committee hearing, senators from both sides of the aisle analyzed when businesses sharing algorithms powered by artificial intelligence cross into illegal anti-competitive conduct.

Chair Amy Klobuchar (D-Minn.) warned that when competitors decide to “delegate their independent pricing decisions to an algorithm, the result is little more than a sophisticated cartel hiding in code.” In his opening remarks, Sen. Mike Lee (R-Utah) emphasized that algorithms aren’t “inherently sinister” and “the devil is in the details”—they can enable or impede competition.

Competition Distortion

The Brookings Institution’s Bill Baer, who led the Department of Justice’s Antitrust Division under President Barack Obama, began his testimony by educating the committee on the two primary ways algorithms can distort competition.

The first scenario involves competitors agreeing to use specific pricing algorithms. The Antitrust Division has been prosecuting these “competitor-to-competitor” agreements for nearly a decade.

In 2015, the Division obtained individual and corporate felony pleas in an international conspiracy to fix the prices of posters sold online. I served as lead attorney on that case during my time with the Justice Department, and Baer is absolutely correct that it was a clear criminal violation under antitrust law.

The second situation is a closer call, Baer said. It occurs when companies use a common vendor to collect “data on supply and demand and recommend pricing or output behaviors that facilitate price coordination.” Baer compared this to “hub-and-spoke conspiracies.”

In antitrust terms, hub-and-spoke conspiracies occur when a company at one level of the supply chain facilitates and joins an unlawful agreement between horizontal companies at a different level, usually by serving as a conduit for exchanging competitive signals and information, The Federal Trade Commission and DOJ have prevailed against Toys “R” Us and Apple Inc., respectively, on a hub-and-spoke charges.

However, Klobuchar, Lee, and Baer all recognized that antitrust law may not reach this second scenario because there’s no “agreement” between independent economic actors. Section 1 of the Sherman Antitrust Act, which governs conduct between companies, requires an agreement that unreasonably restrains trade. Unilateral actions generally don’t trigger a violation, even if they result in lower output or higher prices.

Coordinate Conduct

The distinction between unilateral and coordinated conduct is important to understand. As the US Supreme Court has long recognized, “Congress treated concerted behavior more strictly than unilateral behavior” when enacting the Sherman Act.

This reflects the belief that concerted activity “deprives the marketplace of the independent centers of decisionmaking that competition assumes and demands.” But there can be no concerted action without a “meeting of the minds.”

Federal courts require evidence that “tends to exclude the possibility” of independent actions by defendants. Whether direct or circumstantial, it must reflect a “conscious commitment to a common scheme designed to achieve an unlawful objective.”

Without such evidence, an antitrust action would face heavy odds. Baer wrapped up his testimony by warning that companies can achieve collusive results, absent an agreement, by “writing code that is roughly comparable” and letting AI do the rest. He expressed concern that AI can make coordinated action easier to implement and harder to detect.

The Agreement Is the Code

As in the context of distributed ledger technology, antitrust cartel theory dictates that successful conspirators must solve three challenges. First, they need to agree on terms. For example, they must reach a consensus to charge a fixed price, not pursue each other’s customers, or cut production.

Second, they must ensure everyone is following that agreement. A cartel won’t work if conspirators never implement or follow the agreement.

Third, there must be a way to to punish companies that “cheat.” If one company tries to increase sales by discounting below the agreed-on price, the other conspirators must have some way to make that company pay.

Unless these challenges are solved, the cartel can’t last long. AI and algorithms can instantly observe, synthesize, and respond to vast amounts of sales, purchases, and transaction data.

This could make it easier to know whether a conspirator is cheating on its cartel partners by selling below a fixed price or by producing above an allocated quota. In addition, companies need not meet or even communicate directly to demonstrate that they are complying with the agreement.

That role gets outsourced to the AI. As smoke-filled rooms give way to lines of code, the path to proving an agreement gets even more difficult.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Henry Hauser is antitrust counsel at Perkins Coie. He was previously an antitrust enforcer with the Department of Justice and Federal Trade Commission.

Write for Us: Author Guidelines