AI Personalized Pricing: Is Your Price Set Just for You?

You and your partner order the same groceries, from the same store, at the same hour. Your basket comes to $84.43. Theirs comes to $90.47. Nobody typed in a coupon, and nobody told either of you why.

That scenario isn't hypothetical. It is one of the real baskets a consumer study recorded, and it sits at the center of the fight over AI personalized pricing: software that can set a different price for each shopper instead of one price for everyone. In late October, the Federal Trade Commission ordered nine big companies, from Delta to Walmart, to explain exactly how their prices are set.

This guide covers what the FTC asked for, what the evidence shows so far, how much a hidden price gap can cost a household in a year, and the simple checks you can run before you pay.

What the FTC Just Asked Companies to Hand Over

The FTC used a tool called a Section 6(b) order. It is a study demand, not a lawsuit. It forces a company to answer detailed questions, and no wrongdoing has been found.

According to the FTC's own announcement, the nine companies must document the inputs, tools and procedures they use to set prices and pay. The agency wants to learn whether the price of a product differs from person to person based on loyalty membership or "algorithmic analysis" of personal data. Commissioners approved the orders 2 to 0.

The order also covers workers. Ride-hail and delivery drivers may see pay offers shaped by data such as phone usage or how often they accepted past jobs. The same question runs in both directions: is the number on your screen based on you?

Who got an order

• Airlines: Delta, United and JetBlue
• Ride-hail: Uber and Lyft
• Delivery: DoorDash and Instacart
• Retail: Target and Walmart

Several of them have already denied it publicly. Delta told senators last year it has no fare product that targets customers with individualized prices based on personal data. JetBlue and Instacart have made similar statements. The FTC wants the documents behind those claims.

How AI Personalized Pricing Actually Works

Three different things get lumped together under one scary label. Only one of them is truly about you.

Dynamic pricing moves the price for everyone at once. Concert tickets rise when demand spikes, and an Uber surge hits every rider in the zone. It can feel unfair, but the person next to you sees the same number. Satellite internet works the same way, as we covered in our piece on how Starlink's dynamic pricing changes your bill.

Price testing shows different prices to random groups to learn what sells. Shopper A sees $4.29 and shopper B sees $4.79, chosen by a coin flip. Nothing about you drives it, yet you still pay the difference.

Personalized pricing is the one regulators care most about. Here an algorithm uses data about you to predict the most you will pay, then shows you that number. In a staff study released in January of last year, the FTC said pricing vendors described using signals like your precise location, browsing history, mouse movements on a webpage, and the products you leave unpurchased in a cart. Those were preliminary findings, and the commission split 3 to 2 on releasing them.

Why AI made this easier

Old pricing teams set a few price tiers by hand. A machine learning model can score millions of shoppers in real time and adjust each quote. Delta, for example, told investors on a July earnings call last year that about 3% of its domestic fares were priced with AI from the startup Fetcherr, with a goal of roughly 20%. Delta insists that system prices flights and markets, not individual people.

That distinction is exactly what the FTC now wants proven on paper.

What the Instacart Study Found

The strongest public evidence so far comes from groceries. Consumer Reports, Groundwork Collaborative and More Perfect Union recruited more than 400 shoppers in four U.S. cities. Participants loaded identical carts at the same store, at the same time, and recorded what Instacart charged.

The results, published in the Consumer Reports study, were striking:

• One shopper could pay up to 23% more than another for the same item.
• On staples such as cereal, pasta and peanut butter, the low-to-high gap averaged 13%.
• Whole basket totals varied by about 7% on average.
• The researchers estimated the cost at roughly $1,200 a year for a family of four.

Instacart said the variation came from randomized tests run for retailers, not from personal data. The study did not find prices tracking age, income or race. Then, last December, Instacart announced it was ending all item price tests on its platform.

So the grocery case turned out to be testing, not targeting. For your wallet, the effect was the same.

What a Hidden Price Gap Costs You in a Year

Small percentages hide in plain sight. Run the numbers and they stop looking small.

Start with the real basket from the study. At one Target store in North Canton, Ohio, the same groceries came to five different totals, from $84.43 up to $90.47. The gap between the lowest and highest is $6.04 per order.

• Weekly order: $6.04 × 52 weeks = $314 a year if you land on the high side every time.
• Twice a week: $6.04 × 104 = $628 a year.
• Add rides and flights: a family spending $1,500 a year on ride-hail and delivery fees loses another $105 at a 7% gap.

Random tests should average out across many orders, since you lose some and win some. A personalized price does not average out. If a model has learned that you rarely compare prices, you are on the high side every single time, and that is the scenario the FTC is trying to rule in or out.

Use our inflation calculator to see how an extra few hundred dollars a year compounds into lost buying power over a decade.

How to Tell If a Price Was Set for You

You can't see inside a pricing model. You can still catch one in the act with a few quick comparisons.

1. Compare logged in and logged out. Check the price in your app, then in a private browser window with no account. A big gap is a signal worth noting.
2. Use a second device. Ask someone in your household to load the same item, route or flight at the same moment. Screenshot both.
3. Watch for loyalty-only prices. Member prices are legal and usually disclosed. A price that rises after you join a program is a different story.
4. Check the label in New York. Section 349-a of the state's General Business Law requires the notice "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA" next to such prices. The statute allows civil penalties of up to $1,000 per violation.
5. Report what you find. Send screenshots to the FTC at ReportFraud.ftc.gov, or to your state attorney general.

For flights, timing still matters more than any of these tricks. Our guide to the airfare booking window that saves the most covers when fares usually bottom out.

What Happens Next

The companies have 45 days from receiving their orders to respond. That puts the first answers around mid-December. The FTC can then publish a report, open separate investigations, or do nothing.

Airlines add a twist. The FTC Act largely keeps the agency from regulating airlines directly, which is why it is using a study order here. Consumer protection rules for carriers sit with the Department of Transportation. So even a damning finding about flights would travel through another agency first.

Here's the thing: a disclosure rule may matter more than a ban. Most shoppers can't prove a price was personal. A required label, like New York's, flips the burden onto the seller. Expect more states to copy that model while the federal study runs.

Frequently Asked Questions

Is personalized pricing illegal in the US?

Not as a general rule. Charging different people different prices is often legal unless it discriminates by a protected trait such as race or sex, or it deceives shoppers. New York requires a disclosure when personal data sets a price, and the FTC is now studying whether broader rules are needed.

What is the difference between dynamic pricing and surveillance pricing?

Dynamic pricing changes the price for everyone based on demand, time or supply. Surveillance pricing, another name for personalized pricing, changes the price for you based on data about you, such as your location, browsing or purchase history.

Does Instacart still charge different prices for the same item?

Instacart said last December that it ended all item price tests on its platform. Retailers still set their own prices, and those can differ by store or region, so two nearby stores may not match.

Do airlines use AI to set ticket prices?

Yes, for fares in general. Delta has told investors it uses AI from Fetcherr to price part of its domestic network. Delta, JetBlue and the trade group for U.S. airlines all say they do not set fares for individual passengers based on personal data.

How can I avoid personalized prices online?

Compare the price while logged out or in a private window, check on a second device, and clear cookies before buying something expensive. Screenshots give you evidence if you want to file a complaint.

The Price on Your Screen Should Not Be a Secret

The FTC's orders don't change a single price today. They do something slower and more useful: they force nine of the companies you buy from most to explain, in writing, how their numbers are made.

Until those answers arrive, treat any app price as a starting quote. Spend 30 seconds checking it logged out or on a second phone before a big purchase. If the gap is real, screenshot it and report it, because regulators can only act on the cases people bring them.