A federal antitrust lawsuit filed this week alleges that McDonald's uses an AI pricing engine to set menu prices across U.S. restaurants and to overcharge customers for Big Macs and fries. CNBC's story on the suit ran on October 11, 2026. McDonald's said it does not use AI to judge what an individual customer will pay. It said it gives franchisees tools, research, and recommendations so they can decide.

CNBC's report puts the suit next to a wider shelf of retail software. Kroger said earlier this year that it uses a platform called FlashFood to mark down food near the end of its shelf life and to show those deals in an app. Electronic shelf labels, the small screens that replace paper price tags, are in use at Kroger, Amazon Fresh, Walmart, and Whole Foods, and at Tesco, Morrisons, and Asda in Britain. Revolut has tried a checkout in some coffee shops that charges a customer by recognizing the face. CNBC asked Amazon Fresh, Whole Foods, Tesco, Morrisons, Asda, and Revolut about the AI and had not heard back when the story ran.

The Bank of England wrote about this in April

Clare Lombardelli and Rupal Patel, economists at the Bank of England, wrote in April that better software makes prices move more often and makes them more specific to a person. They described the endpoint as charging close to the most a given shopper will pay, and they called that perfect price discrimination. A consumer-price index, they said, assumes a representative price that moves slowly and together. When each shopper sees a different sticker, that sample gets harder to read, and households can end up with different inflation rates. A recent print had inflation at 3.4 percent against wage gains of 3.1 percent.

A personal price is a different system from a market price

Miroslava Marinova, a senior lecturer in commercial law at the University of East London, told CNBC that changing a price because demand, the clock, capacity, or a rival moved is an old practice. Airlines, hotels, and ride-hailing have done it for years. Her distinction is the second system: a price built from what the store knows about the person. Transaction history, browsing, location, and what someone usually buys are the inputs she listed. On Wednesday, Sainsbury's released SmartLists, which builds a shopping list from a photo or a typed meal. Marinova said that once a retailer mixes market data with that personal file, the line between the two kinds of pricing gets thin.

New York, Maryland, New Jersey, and Connecticut have statutes

New York requires most businesses that use personal data to set a price to say so clearly. Maryland has limited food retailers and delivery services in using personalized prices to charge more for some food. New Jersey and Connecticut have passed measures aimed at what the statutes call surveillance pricing. Walmart and Kroger have said in recent years that their tools run the operation, and that they do not set a surge price for one named customer. A consumer outlook index has fallen to 47.8 as inflation views darkened.

The Bank of England economists added that personalized prices splinter what a household experiences, so a single inflation average can stop matching the sticker any one shopper paid. Marinova said a lower price for some people is the other result of the same file, the case where the data produces a discount. The risk she pairs with that discount is a shopper who cannot tell the market's price from a price built out of browsing, location, and past purchases. Comparing two stores, she said, stops being a check of one shared sticker, and walking away puts less pressure on the retailer.