DOJ antitrust official warns AI pricing systems could trigger criminal collusion cases

Acting Deputy Assistant Attorney General Daniel Glad warned Thursday that artificial intelligence and algorithmic pricing tools will not shield companies or executives from criminal antitrust liability, declaring that "software cannot launder collusion" in a major policy speech delivered at the Antitrust West Coast Conference in San Francisco.

In remarks titled "Old Crime, New Code," Glad said the Justice Department's Antitrust Division is intensifying scrutiny of pricing algorithms, large language models and AI-driven business tools that could facilitate unlawful coordination among competitors. He emphasized that traditional antitrust principles apply regardless of whether companies communicate through meetings, text messages or automated software systems.

"The tools change," Glad said. "The rule against collusion does not."

The speech, delivered at Informa Connect's Antitrust West Coast conference, drew a who's who of antitrust regulators, including Daniel Graulich, chief of staff to FTC Commissioner Mark Meador, and Paula Blizzard, senior assistant attorney general in the California Department of Justice's Antitrust Section.

Glad, according to a transcript of his speech, offered one of the clearest indications yet that federal prosecutors are preparing for potential criminal cases involving AI-assisted pricing and algorithmic coordination. Glad pointed to the Justice Department's recent settlement with RealPage, a property management software company accused of facilitating unlawful information sharing among landlords, as a blueprint for future enforcement. Under the 2025 consent decree, RealPage agreed to restrict its use of nonpublic rental data and submit to court monitoring.

Glad stressed that while the RealPage matter was resolved civilly, algorithmic conduct remains subject to criminal prosecution where evidence supports an agreement among competitors to fix prices or allocate markets. He said the key legal issue remains whether competing firms knowingly shared sensitive data or relied on common systems in ways that displaced independent decision-making.

The speech also highlighted the department's growing reliance on digital evidence and whistleblowers. Glad said automated systems often create more discoverable records than traditional conspiracies, including software logs, data inputs and algorithmic changes. He noted that the department recently awarded its first $1 million whistleblower payment in a bid-rigging investigation tied to online vehicle auctions.

Glad warned companies that the combination of the Antitrust Division's leniency program and new whistleblower incentives creates a "second race" in which employees, engineers, consultants and data scientists may report suspected misconduct before corporations self-disclose violations.

He urged companies using AI and pricing software to strengthen compliance programs and ensure antitrust review is integrated into broader AI governance efforts.

"'The model did it' is not compliance," Glad said. "It is the beginning of the next question."