Walmart recently published a list of things it promises not to do to its customers.

“We price the product, not the person,” Walmart CEO John Furner announced in a September letter to customers. “Your income, shopping history, urgency or what we think you could pay won’t change the price,” he promised. Nor, he said, would the retailer’s AI shopping assistant, Sparky, use information customers shared to raise prices or conceal cheaper options.

This was nice to hear. I hadn’t previously considered that telling a computer I needed toilet paper might result in an emergency toilet-paper rate. There is something faintly unnerving about the specificity of a corporate reassurance, particularly when it introduces several possibilities you hadn’t thought to worry about.

It made me preemptively nostalgic for the little paper price tag beneath a box of cereal. There’s something reassuringly stupid about it. It tells you what the cereal costs. It doesn’t know how much money you make, whether you’re in a hurry or what the temperature is outside. Unless an employee comes along with a new sticker, the person reaching for the same box ten minutes later sees the same number. A modest achievement, perhaps, but one I’m growing increasingly fond of.

The appeal isn’t really the paper. It’s the transparency of the arrangement. As retailers gather more information about their customers and acquire increasingly sophisticated pricing tools, shoppers risk knowing less about why they’re paying what they’re paying. The problem isn’t simply higher grocery bills. It’s the possibility of a transaction in which one side can examine the other in exquisite detail, while the other can barely see the rules.

What prompted the letter from Furner is that Walmart is expanding its use of electronic shelf labels, little screens that let employees update prices digitally from a central computer instead of replacing stickers by hand. The change makes sense. Furner points out that updating paper tags is tedious work, and he’d rather employees spend their afternoons helping customers than tending to a small forest of adhesive rectangles. Digital labels can also help ensure the price on the shelf matches the price at checkout.

And electronic labels aren’t, in themselves, surge pricing. Walmart says it doesn’t use them to change prices according to who is shopping or the time of day. But efficiency is a different question from trust. Making a price easier to change isn’t necessarily a problem. Making it harder to understand how that price was reached might be.

---

Want more great food writing and recipes? Sign up for Salon’s free food newsletter, The Bite.

---

We’ve already accepted a surprisingly elaborate set of conditions for saving money on groceries. Download the app. Make an account. Remember the password. Clip the digital coupon before you leave home. After handing over our personal information and completing several minor administrative tasks, we’re permitted to save 75 cents on yogurt. The store, meanwhile, gets a rather detailed account of our preferences. What happens when that bargain runs in reverse?

Last December, Consumer Reports and Groundwork Collaborative found that shoppers using Instacart were sometimes shown different prices for identical groceries from the same store at the same time. At one Safeway in Washington, D.C., an 18-ounce box of Signature SELECT Corn Flakes appeared for $2.99, $3.49 and $3.69, depending on whose screen you were looking at. 

Instacart said the shoppers were assigned to pricing experiments randomly, not sorted by income, demographics or browsing history. It ended the tests after the investigation. That distinction matters, especially because the research didn’t establish that anyone’s personal habits were being used against them. But it demonstrated how quietly prices can diverge when the shopper sees only the number assigned to her. You can’t question a difference you have no way of seeing.

I buy a lot of black beans. Enough, presumably, that a supermarket’s software could identify me as a woman with firm opinions about legumes. If that earns me a coupon, wonderful. I’m considerably less enthusiastic about the possibility of an algorithm deciding that my apparent dependence on black beans means I could stand to pay another 30 cents per can. It’s one thing to recognize a loyal customer. It’s another to recognize a captive audience.

Other pricing algorithms raise different questions (McDonald’s, for instance, faces a lawsuit alleging AI-assisted price coordination among franchises, which the company denies). Prices have always fluctuated; coupons have always rewarded some shoppers more than others. Nor should computers be banished from the business of deciding what things cost. But the supermarket is not an airline, and dinner should not require the same vigilance as booking a flight.

Walmart’s promise is worth taking seriously. The company has drawn a clear line around how it says it will use customer information, and that’s preferable to leaving shoppers guessing. But it’s a corporate commitment, not a rule every retailer must follow. And the fact that a supermarket now feels obliged to assure us it won’t charge extra because we seem hurried or financially comfortable suggests how strange our expectations have become.

Perhaps it’s naive to want a supermarket not to use what it knows about me against me. Still, I don’t need the price tag to know me. I need it to tell me what the cereal costs.

The post The strange new math behind the price of your groceries appeared first on Salon.com.

As Amazon blocks AI agents from shopping on its site, Walmart is taking the opposite approach, introducing new technology that lets AI agents make purchases on customers' behalf, with controls over spending and transaction approvals.

Flipkart has expanded access to its food-delivery service to thousands of employees ahead of its public launch.