After scrutiny over algorithmic pricing, two of the country’s largest consumer brands are trying to define what they will — and will not — let technology do with prices.

Walmart and McDonald’s are responding to growing concern over how artificial intelligence could influence what consumers pay, but they are drawing the line in different places. Walmart has publicly promised not to use personal information such as income, shopping history or urgency to charge individual customers more, while McDonald’s says its AI pricing tools do not automatically set menu prices or adjust prices based on individual customer behavior. The statements follow heightened scrutiny of McDonald’s use of an AI-powered system that analyzes transaction data across roughly 14,000 U.S. restaurants and recommends what the company considers an “optimal price.” Three franchisees told Reuters that the system had widened some price differences between nearby restaurants, while McDonald’s maintains that franchisees retain final control over pricing decisions. Pasted text

The Argument Is Moving From Capability to Boundaries

The technology itself is no longer the only issue. Retailers already have access to detailed purchasing data, digital shelf labels, loyalty histories, local demand information and software capable of changing prices quickly. The more important question is what companies decide those systems are allowed to do. Walmart CEO John Furner recently framed the company’s position with a simple promise: “We price the product, not the person.” According to material shared by the company and industry observers, Walmart says income, purchase history and urgency will not be used to increase the price a specific customer sees, even as the retailer expands digital shelf labels and AI-assisted shopping tools. Pasted text

SSC has been following this question beyond retail. In Cities Are Using Dynamic Pricing for Parking. Public Space Is Becoming Variable Cost., we examined how cities are using demand-based parking rates to manage scarce curb space. The policy may improve efficiency, but it also creates a distribution problem: people with flexible schedules can respond to changing prices, while workers, caregivers and others with less control over when they travel often simply absorb the higher cost.

That pattern became even clearer in SSC’s more recent reporting on New York’s congestion-pricing system. Researchers found that the same $6.75 price difference changed behavior at only two of four toll boundaries. Visits rose 5.9% around the weekday 5 a.m. cutoff and 9.2% around the weekend 9 p.m. cutoff, while the other two transition points produced no statistically meaningful change. The deciding factor was not simply the size of the price signal. It was whether people had enough flexibility to change their schedules.

McDonald’s Still Uses AI to Recommend What a Market May Bear

McDonald’s denial of personalized or automatic pricing does not mean AI is absent from the process. The company’s system analyzes transaction data and local market conditions to recommend prices to franchisees. Reuters reporting cited in the supplied material found that a company-run McDonald’s in Fresno sold a Big Mac for $5.69, while another company-operated location about two miles away charged $6.89 — a 21% difference. Reuters did not establish that AI caused that specific gap, and McDonald’s has emphasized that local pricing differences existed long before artificial intelligence. Pasted text

McDonald’s has also pushed back against descriptions of its system as dynamic pricing. Jon Banner, the company’s global chief impact officer, wrote that AI does not set menu prices, make automatic price changes or determine what to charge individual customers. He said franchisees make final pricing decisions based on their businesses and local markets, while corporate tools are designed to provide recommendations and information. Pasted text

The connection to SSC’s earlier reporting is useful because it shows the same mechanism appearing in different settings. Cities change prices to influence behavior. New York’s toll data shows that the response depends on whether people can actually adjust. McDonald’s is using transaction data to estimate what a particular market may be willing to absorb. In one case, the price tests behavior. In the other, behavior helps inform the price.

Walmart Is Making Trust Part of the Product

Walmart’s response goes further because it is making a public commitment about what its data will not be used for. The company is rolling out digital shelf labels across thousands of stores and operates an AI shopping assistant capable of learning from customer behavior. Those tools give Walmart greater technical flexibility, but Furner’s promise effectively says the company will not use that capability to charge one shopper more for the same product because an algorithm concludes that person is willing or able to pay more. Pasted text

That is becoming a competitive issue as much as a regulatory one. Some states are beginning to legislate around personalized pricing, while federal regulators have explored disclosure rules. But company policies can move faster than law. When one major retailer publicly says which forms of algorithmic pricing it will reject, competitors face pressure to explain their own boundaries.

The distinction among dynamic pricing, personalized pricing and algorithmic pricing is therefore becoming more important. Dynamic pricing generally changes prices based on conditions such as demand, inventory, competition or time. Personalized pricing changes what a particular customer pays based on information about that person. Algorithmic pricing describes the use of software to recommend or help determine prices, whether or not the final number changes automatically. Consumers may view all three as the same thing if they cannot see how the number was produced, but companies are beginning to treat those categories as lines they need to define publicly.

The New Pricing Question Is Who the Algorithm Is Optimizing For

Businesses have always adjusted prices by market, location and demand. AI makes those judgments faster, more granular and potentially more predictive. A company can increasingly estimate not only what an item costs to sell, but what customers in a particular market are likely to accept.

That creates a trust problem before it creates a legal one. Consumers may accept that groceries cost more in one city than another or that restaurants respond to local operating costs. They may be less comfortable with the idea that the same data used to personalize recommendations could eventually be used to identify how much more they might tolerate paying.

SSC has now followed this pricing logic through parking, congestion tolls and fast food. The common feature is not simply that prices move. It is that data is making those prices more responsive to behavior, location and demand. For consumers, the harder question is whether they have any meaningful ability to respond in return. Walmart and McDonald’s are now drawing public boundaries around what their systems will do because companies are beginning to understand that the trust problem starts before a price becomes personalized. It starts when people no longer know what is shaping the number in front of them.