September 2, 2026

The Price Is Whatever the Machine Thinks You'll Pay

A traveler trying to get to a funeral watched a fare jump $230 overnight and was told to clear his browser history. California, New York and the FTC are all moving on "surveillance pricing" this week. The catch is the economists don't agree it's bad for you.

The facts:

• A 2025 lawsuit against JetBlue cited a customer's post that a fare had risen $230 in a day while he tried to reach a funeral; a JetBlue rep replied suggesting he clear his browser history and try again; JetBlue denies using personal data to price tickets (Los Angeles Times)
• California's Assembly Bill 2564 would ban "surveillance pricing," defined as using personal data to set a price by any technology including AI; the Assembly passed it in May and the Senate took it up Sunday against a Monday midnight deadline (Los Angeles Times)
• New Jersey has already passed a ban, a New York bill is waiting for the governor's signature, and the FTC said this month it will go after companies that don't disclose personalized pricing (Los Angeles Times)
• In 2022 California settled with Target for $5 million over charging different app prices for the same item depending on where the shopper was standing (Los Angeles Times)
• An FTC study of eight companies, including Mastercard and JPMorgan Chase, found pricing systems that log whether you highlighted a product name and how far down a page you scrolled; one example: a shopper profiled as a new parent gets shown pricier thermometers first, and choosing faster shipping can flag desperation for future pricing (Los Angeles Times)
• Consumer Reports found Kroger built demographic profiles to tailor discounts, and in one case most of what it believed about a shopper was wrong (Los Angeles Times)
• 76% of Americans say charging different prices based on personal data is unfair; researchers say there's little hard data on how widespread it is (Groundwork Collaborative, via Los Angeles Times)
• Some economists argue personalized pricing can mean lower prices for price-sensitive shoppers, and warn a ban could remove those discounts (Los Angeles Times)
• The FTC finalized $930,000 in penalties against Cox Media Group and two partners for advertising an "Active Listening" service that supposedly targeted ads from conversations picked up by smart devices; the service didn't actually use voice data, and the FTC noted that if it had, that would have been illegal too (FTC)
• A Target worker was filmed by customers wearing Meta smart glasses who kept asking for a price check to get a reaction; saying "Meta, stop recording" to someone else's glasses does nothing; the video got hundreds of thousands of views (The Verge, via Jacobin)
• ChatGPT's new plugin can search years of a Mac user's text messages; the person who installs it consents, and nobody else in those conversations does (Fortune)

For most of history the price of a thing depended on who was asking. The seller looked at your sandals. One price for everyone was a Quaker idea, and it won because it felt fair, and because a price tag is faster than a negotiation. The tag is what's going away.

What replaced the sandals is the profile. Where you live, what you searched, whether you highlighted the product name, how far you scrolled, whether you paid for two-day shipping last time. The FTC's own report describes a system that reads a new parent shopping for a thermometer and shows the expensive ones first. Not because the thermometer costs more. Because you might pay more.

That's the part that lands. The price isn't a fact about the product anymore. It's a guess about you.

the case for it

Be fair to the other side, because it's not crazy. If a store knows you can't pay much, personalized pricing can mean a discount you'd never have gotten under one price for all. The airline that charges the desperate traveler more may also be the airline that charges the flexible student less. Economists who study this say bans could wipe out the low end along with the high end, and that the data on how common the practice actually is comes mostly from anecdotes.

The Kroger example is the counterweight. The store built a profile of a shopper to tailor his discounts, and most of what it believed about him, his education, how many kids he had, was wrong. The machine doesn't have to be right to set your price. It only has to be confident.

the week's smaller stories are the same story

The FTC fined a media company for claiming it could listen to your conversations through your devices and sell ads against them. The punchline in the order: it couldn't. It made the claim up to sell the service. Which tells you what the customers wanted to buy.

A Target worker got filmed by two customers in smart glasses who kept asking the price of a $20 item to get a reaction. The worker tried saying "Meta, stop recording" at another person's face. That command only works on your own glasses.

And ChatGPT can now read a Mac user's text messages back for years, on request. The person who installs it agrees to that. The person who sent the texts never hears about it.

None of these is a scandal on its own. Together they're a direction: every interaction gets recorded, every record gets scored, and the score comes back to you as a price, a discount, an ad, or a video with your face in it. Fairness, the thing the Quakers were selling, was a rule that the seller didn't get to know too much.

California had until midnight Monday to decide whether to write that rule down. New York's version is on the governor's desk. Either way, the seller already knows what you're wearing.

Sources: Los Angeles Times, Federal Trade Commission, The Verge via Jacobin, Fortune.