The AI deciding what you pay

Here's something AI quietly made cheap that you should know about: charging you, in particular, a different price than the person next to you — for the exact same thing, based on your personal data. It's called surveillance pricing, and as of this month it's squarely in regulators' sights. New York just moved against AI-assisted surveillance pricing in its latest session, on top of a disclosure law already on the books, and it's far from alone.

Most people have never heard the term, so let me make it plain — because once you see it, you can't unsee it on your next checkout screen.

What it actually is (and isn't)

Surveillance pricing is when a company uses your personal information — location, browsing history, device, inferred income or family size, even your mouse movements and what you left in your cart — to guess the highest price you personally will pay, and then shows you that price.

It's important to separate this from ordinary dynamic pricing, which is fine and familiar: prices that move with the market — surge pricing when demand is high, cheaper flights in the off-season. That responds to the world. Surveillance pricing responds to you. The regulators draw exactly this line: dynamic pricing reacts to supply and demand; personalized pricing reacts to the characteristics of the individual.

The examples from the FTC's own study are unsettling. A shopper profiled as a new parent might be shown higher-priced baby thermometers first. An investigation into one grocery-delivery app found its pricing could cost some customers around $1,200 more a year. The whole point is to find the number you, specifically, won't walk away from — which is the opposite of a fair market price.

Why this is the moment

AI is what made this practical at scale. Inferring each shopper's willingness to pay used to be hard; now a model can do it from the same data trail that powers "personalized" recommendations. So the regulation arrived fast and broad. New York's Algorithmic Pricing Disclosure Act already forces businesses to literally print "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA." Maryland went further and banned using surveillance data to raise prices in some categories. By spring, more than 40 bills across two dozen states were taking aim, with the FTC and Congress investigating travel and platform companies.

What it means for you

Two practical reads, depending on which side of the checkout you're on.

If you're a shopper: just knowing this exists changes how you read prices. The same flight, cart, or subscription can cost different people different amounts on purpose — not because of the market, but because of what the seller inferred about you. That's why comparing across accounts, logging out, or clearing data sometimes moves the price. The asymmetry is the point: they know a lot about you, and you know nothing about how it set your number. The new disclosure laws exist precisely to close that gap.

If you build products: there's a line here worth internalizing before a law makes you. AI gives you enough knowledge of each customer to do two very different things — to serve them better, or to squeeze them harder. The same data that personalizes a useful experience can personalize the price you extract. Dynamic-by-market pricing is fine and legal. Personalized-by-the-person's-data pricing is becoming illegal in more places every month, and even where it's legal, it's a trust bomb: the moment customers learn their price was set from their data, they don't feel served, they feel hunted.

The bottom line

Surveillance pricing is a clean example of the broader thing AI does to every business: it makes it cheap to know each customer intimately — and intimacy can be used to help or to exploit. The regulation sweeping the states is really just drawing that line in law: you can use AI to understand your customer, but not to turn that understanding against them at the cash register.

So as a buyer, know the practice exists, because it's already shaping the prices you see. And as a builder, decide which side of "serve versus squeeze" you're on before the law decides for you — because the same model that lets you treat a customer better lets you treat them worse, and customers, regulators, and increasingly the courts can tell the difference. The cheapest trust you'll ever keep is the price you didn't quietly personalize.