Companies of all sizes and across every sector, including banking and financial services, have been scrambling to integrate artificial intelligence into their business models and day-to-day workflows. The possibilities offered by AI have seemingly driven companies and employees into an adoption frenzy. But the rush to drink from this digital fountain of youth is not without risk, as a recent case shows.

Imagine that sales teams at competing companies use the same AI-driven pricing model. Each team uploads their company’s confidential price lists and rates, planned discounts, forecasts, and customer population data, then asks the model to recommend pricing. The model provider retains the uploaded content, uses that content to improve its AI-driven pricing services, and allows the model to account for what it learns from one user to influence the pricing guidance given to a user at a competing company. The companies then adhere to the model’s pricing recommendations, even though those  terms may be different than what each company would have otherwise adopted. Could the companies later be dragged into court based on allegations of anticompetitive conduct?

The U.S. Court of Appeals for the Third Circuit has answered that question with a “yes.”[1] On July 29, 2026, the Third Circuit held that when an AI-driven “algorithm is in effect collecting non-public commercial information from [competitors] and utilizing the collective pot of data to ‘suggest’ prices to each [competitor],” that “surely raise[s] a plausible inference of collusion under Section 1 of the Sherman Act.”[2] The court’s holding, discussed further below, should serve as a cautionary tale for businesses of the risks that should be considered when deciding whether (and how) to use AI to compete in the marketplace.

The Third Circuit’s Decision in Cornish-Adebiyi v. Caesars Ent. Inc.

Casino-hotel guests brought a putative class action against several Atlantic City casino hotels, alleging a horizontal price-fixing conspiracy under § 1 of the Sherman Act. The plaintiffs claimed the defendant casino hotels and their shared algorithmic software provider, Cendyn, conspired to fix hotel room prices.

Each casino hotel fed its room pricing and occupancy data into Cendyn’s AI-driven dynamic pricing program called “Rainmaker,” which then processed that data—along with  data from competing casino hotels—and generated suggested room rates that were then uploaded into the participants’ systems. Although hotels allegedly retained ultimate pricing authority, deviations required special override permissions, and hotels followed Rainmaker’s recommendations 90% of the time.

The plaintiffs alleged this was a stark departure from the casino hotels’ historical practice of offering deeply discounted room rates (such as to draw gamblers onto the casino floor). The plaintiffs alleged that this arrangement replaced historically independent pricing, enabled hotels to avoid competition in undercutting one another, and led to rising room rates despite declining occupancy.

On appeal, the Third Circuit examined how AI-powered dynamic pricing algorithms can facilitate anticompetitive behavior. The court acknowledged that there is nothing inherently anticompetitive about using algorithms. However, it emphasized that AI software can facilitate collusion by enabling competitors to coordinate prices and share information without ever directly communicating with one another, and that real-time price monitoring enables “cartels” to more effectively police each other’s pricing behavior. The court noted that historically, collusion was hindered by communication gaps and enforcement costs, but that today’s AI algorithms have the capacity to bridge those gaps—making widespread coordination possible. Thus, the court held that the plaintiffs plausibly alleged the hotel casinos violated the antitrust laws.

Invoking a notable analogy from former FTC Acting Chair Maureen Ohlhausen, the court summarized: if it is not permissible for a person named Bob to collect confidential pricing strategy information from all market participants and then tell each one how to price, it is probably not permissible for an algorithm to do it either.

Practical Takeaways

The Third Circuit’s decision illustrates why companies, including banks and financial institutions, must be conscious of the risks of developing and using AI-driven pricing models and algorithms. Indeed, the message to companies is clear—using AI to automate competitive information sharing and coordinated decision-making is subject to antitrust scrutiny. Those looking to account for those risks should avoid over-use of information-sharing AI platforms; document independent decision-making; limit competitors’ data inputs on the platform; and audit AI tools for synchronized pricing, reduced competition, and other potentially collusive outcomes.

Banks and other financial institutions should be particularly careful when deploying AI tools that influence pricing, rates, fees, discounts, or other competitive terms. Before providing proprietary pricing information, customer data, forecasts, or other competitively sensitive information to a third-party AI platform, financial institutions should understand how the provider retains, uses, and combines that information—including whether it may be used to train the model or inform recommendations provided to competitors. Institutions should also consider contractual and technical safeguards that prevent the commingling of competitively sensitive data, maintain meaningful oversight over pricing decisions, and document the independent business reasons supporting departures from—or acceptance of—algorithmic recommendations. Put simply, financial institutions should treat an AI pricing platform not merely as another piece of software, but one that warrants special antitrust, compliance, and data-governance review.

[1] Cornish-Adebiyi v. Caesars Ent., Inc., — F.4th —-, 2026 WL 2182291 (3d Cir. July 29, 2026).
[2] Id. at *12.

August 24, 2026