New U.S. DOJ AI anti-trust guidance
For corporate compliance officers struggling to define your role in AI governance at your employer, the U.S. Justice Department just did you a favor. One of the department’s top antitrust officials recently outlined his vision for when abuse of AI systems might lead to a regulatory enforcement action – and how compliance officers could help companies avoid that predicament.
That official is Daniel Glad, head of criminal antitrust enforcement at the Justice Department. Glad delivered a speech on May 14 that raised the specter of “algorithmic antitrust enforcement,” such as when a group of companies all use a single AI system to help them set prices for the goods they sell to consumers.
Is AI-assisted collaboration like that inherently illegal? No, Glad stressed. But could companies misuse those AI systems in a way that does lead to antitrust violations and regulatory enforcement?
Glad was just as clear on that second question, too: Yes, absolutely.
So, oversight of how companies use AI systems will be critical. That’s where compliance officers can enter the chat.
What is ‘algorithmic antitrust risk,’ anyway?
We should start by doing a close read of what Glad said. Algorithmic antitrust risk arises when multiple companies put confidential information into the same AI tool, which then gives those multiple companies a single price (or set of prices) they impose upon consumers.
“Where the evidence shows that competitors used a system – any system – to replace independent decision-making with shared competitive intelligence, we will treat that as what it is,” Glad said.
That point is nothing new in the antitrust world. If a group of competitors met in a bar during an industry conference to decide on prices they’d all use, or did the same via conference call or group text message, everybody would agree that yes, that’s an antitrust violation. So why should using an AI system for that purpose be any different?
The Justice Department has already put that theory to work in the courtroom. In 2024, federal prosecutors brought a civil lawsuit against six large apartment management companies for using a single AI software system to set rent prices. Each company put its confidential information (rents, tenant payment histories, renewal rates, and so forth) into the software tool, which processed all that data and then told the companies what rents they should impose on the apartments they owned. Several of the landlord businesses (and the software firm that ran the AI tool, which was also sued) have settled their cases so far.
Glad talked about that case at length in his speech and expressly said: “It does not ban the software. It does not ban algorithmic pricing generally. It targets the ingestion of non-public competitor data and the granular reporting of outputs back to competitors.”
Numerous industries rely on so-called “reference prices” to help them understand the specific prices they might charge consumers. The Justice Department is reportedly investigating both the egg industry and the meatpacking industry for price fixing.
So yes, the risk is out there, and it’s real – but the risk isn’t about the tool itself. The risk arises from how your company uses the tool.
Assessing and managing the risk
The first question for compliance officers is whether your management team even appreciates this risk. They might not; they might decide the issue is so abstract and unlikely that you don’t need to play a role in oversight.
That might be a mistake. As Glad noted in his speech, algorithmic antitrust misconduct is likely to be more visible than traditional antitrust schemes, not less.
That is, when a few sales executives conspire over drinks to rig prices, only that small group knows what’s afoot. Use AI to do the same, and the scheme could be visible to a much wider group of people: the software engineers who built the tool, the data scientists who trained the model, and account managers who sell the software to customers, to name just a few.
So, the wiser course of action would be to include the compliance officer in a risk assessment of how your sales team wants to use the tool. (Really, the compliance officer should participate in the risk assessment for any AI tool your company wants to use; but we’re sticking with antitrust issues today.)
And if you do have antitrust risks, what then? Compliance officers could pull any of several levers to mitigate the risk, all of them familiar:
- Policies and procedures. Develop a clear policy for how your AI pricing tools should be used. Consider the documentation and access controls you’d want to ensure the correct people use the tool correctly.
- Training. Employees need to be trained on your antitrust and AI usage policies. Otherwise your policies are only worth the paper they’re printed on.
- Due diligence, testing, and monitoring. Most companies now use AI systems provided by a third party. Compliance officers should review and test those tools (probably with help from the IT team) to understand how the tool works and whether it poses any compliance risks, given how employees want to use it.
- Internal reporting. Glad’s other big message in his speech was that the Antitrust Division is eager to have whistleblowers report schemes directly to prosecutors. If a whistleblower does, your company’s own chance to self-report and receive cooperation credit disappears. So yet again, it’s crucial for companies to cultivate a strong culture of internal reporting and to follow up on those reports when they come to you.
The bottom line is that companies can’t hide behind AI to shield themselves from antitrust risk (or any other compliance risk, for that matter). Corporate misconduct is still fundamentally a matter of human misbehavior, and companies need strong compliance program capabilities to keep that human behavior on the right side of the law.
That’s not going to change no matter how impressive AI is.
How NAVEX helps
NAVEX solutions are designed to help organizations operate ethically and responsibly. Policy & Procedure Management enables you to draft, distribute and track critical AI policies. Ethics & Compliance Training is legally vetted and compliant with U.S. state and global regulations, and is continually updated to keep pace with the fast-moving changes in artificial intelligence. Third-Party Screening & Monitoring helps ensure your partners are compliant with relevant laws and that you have insight into any red flags before onboarding. Finally, Whistleblowing & Incident Management supports your employees, customers and third parties as a channel to report concerns and a case management system to track reports through the case life cycle and resolution.
The conversation about AI and its uses, implications and governance is ongoing and fast-evolving. Join NAVEX on July 15th as we explore this topic more and discuss how Compliance teams can ethically and responsibly use AI in their programs.
