Las Vegas Ruling Offers Roadmap for AI Clashes With Antitrust

A Nevada judge’s dismissal of a price-fixing suit against major Las Vegas hotels is rippling across the country as courts wrestle with the question of how antitrust laws intersect with AI-powered information systems.

Chief Judge Miranda Du of the US District Court for the District of Nevada earlier this month dismissed a proposed class action alleging the hotel-casinos shared algorithms to set room rates, saying there was no evidence of a tacit price-fixing agreement, and that hotels could reject the machine-suggested rates.

Her ruling was among the first to offer clear guides for when AI decisions might cross into illegal conduct. Defendants in at least three other antitrust suits quickly seized upon it, asking that the rationale be considered in their cases.

“It’s a silver bullet for almost all these cases,” said Ken Racowski, a litigation attorney at Holland & Knight LLP, which has defended clients in similar algorithmic price-fixing suits. “If you’re actively litigating these cases, I think it provides a lot of ammunition and fuel for arguments for your clients.”

The Las Vegas case focused on accusations that defendants including Caesars Entertainment Inc., Wynn Resorts Holdings LLC, and Treasure Island LLC collectively used pricing algorithms built by the Rainmaker Group, a subsidiary of tech company Cendyn Group LLC, to inflate hotel prices in violation of antitrust laws.

In a separate case involving pricing algorithms, Yardi Systems Inc. and multifamily property owners cited Du’s ruling as a “supplemental authority” to combat accusations that they conspired to inflate the price of rental properties nationwide. Revenue-management system RealPage Inc. and more than a dozen of Washington’s largest landlords also filed her ruling last week in a pending case from the District of Columbia’s attorney general accusing them of using algorithms to unlawfully inflate rents.

“Judges would certainly react to the kind of clear steps she went through,” said Kathleen Bradish, vice president and director of legal advocacy for the American Antitrust Institute, referring to Du’s opinion.

Attorneys defending Caesars Entertainment and other hotels in a similar case focused on Atlantic City wrote to New Jersey district judges last week to flag Du’s opinion.

“That’s the next shoe to drop, how those three courts decide a motion to dismiss,” Racowski said.

Novel Theory Roadmap

The cases come as more companies across industries rely on artificial intelligence to power business decisions, including to assess risks and identify trends, raising novel legal questions.

William Kovacic, a former Federal Trade Commission chair and current law professor at George Washington University, said the competing cases point to early efforts by plaintiffs to show that using algorithmic price-fixing facilitates collusion on prices.

“We are now seeing a handful of decisions that wrestle with when the adoption of these tools can be deemed to be collective action,” he said.

Du’s ruling against plaintiffs contrasts with a competing decision from Judge Waverly Crenshaw of the US District Court for the Middle District of Tennessee. Crenshaw last year denied a motion to dismiss a case against RealPage after finding evidence landlords provided the revenue-management company with proprietary commercial data, “knowing that RealPage would require the same from its horizontal competitors and use all of that data to recommend rental prices to its competitors.”

However, the RealPage case included allegations that competitors exchanged confidential information through the algorithm, while the hotel price-fixing case didn’t.

The Las Vegas hotel price-fixing case “remains a relatively novel antitrust theory premised on algorithmic pricing going in search of factual allegations that could support it,” Du said in her ruling. She rejected the plaintiffs’ theory pointing to an alleged deal among the casino-hotels to adopt prices suggested by the algorithm. The hotels didn’t use Cendyn’s software to exchange confidential data, were free to reject pricing recommendations, and started using the software at various times, she said.

The use of machine learning also can’t be used to infer an agreement, Du said, noting that the technology is no different than an attorney “improving her skills over time with the benefit of experience and access to confidential client information she gains with each client engagement.”

Bradish said it’s always a “little dangerous” to analogize machine learning to humans. But that analogy helped Du make the distinction between sharing confidential information, and just using it to learn and improve, she said.

“The analogy tried to make it understandable, and the judge certainly thought that was the case,” Bradish said.

The Las Vegas case is Gibson v. Cendyn Group LLC, D. Nev., No. 2:23-cv-00140.