Is AI raising gas prices? Drivers say yes – and sue.

California drivers are suing the largest gas station operators for compensation for overcharges. The defendants include BP, Marathon Petroleum, 7-Eleven, Speedway, Circle K, Walmart, Sam's Club, Albertsons, EG America, TravelCenters of America, and software provider Kalibrate.

The basis for compensation? A violation of competition law through coordinated price-setting. However, the coordination in pricing policy did not happen behind closed doors in a boardroom. The coordination was algorithmic, carried out through a shared pricing platform.

How was the system supposed to work according to the lawsuit? Kalibrate Fuel Pricing is an algorithmic pricing tool connected directly to fuel pumps and price signs at gas stations. According to the complaint, Kalibrate's marketing pitch is simple: in exchange for handing over fuel pricing decisions to Kalibrate Fuel Pricing, the company promises its customers higher margins and profits from fuel sales. Kalibrate's promotional materials state that it achieves these higher prices by collecting data on all competitors in a given area and then applying a pricing algorithm that coordinates users of the platform so that they do not compete with each other on price.

The plaintiffs calculate that a 1-cent increase in gas prices costs California drivers $134 million per year. The average surcharge—and thus, in the plaintiffs' view, the harm—caused by algorithmic pricing is 6 cents per gallon of gasoline. However, the price increase is much more pronounced in areas with a high concentration of gas stations that have deployed Kalibrate Fuel Pricing, where it can reach up to 30 cents per gallon. The defendants operate over 1,700 stations in California.

This lawsuit further illustrates the importance of considering competition law when deploying AI tools—that is, the very topics we discussed last Friday at a webinar dedicated to the risks of AI in the context of competition law. Among other things, participants clarified that algorithmic pricing and the use of AI tools for competitor benchmarking are not automatically in conflict with competition law. However, to ensure the legitimate use of these tools, it is necessary to focus on and properly document the manner of their deployment and use, as well as their training and inference datasets, the design of the algorithm and its technical functioning, and the way human oversight and intervention are exercised. If you missed the webinar, reach out to us—we will be happy to send you the recording.