AI is watching your spending and setting your prices accordingly. Lawmakers want to stop it
In 2025, a New York resident filed a lawsuit against JetBlue Airlines.
The lawsuit included a screenshot of a tweet from X. In it, a user complained to JetBlue that they were trying to “make it to a funeral” but that flight prices had risen $230 in just one day. A JetBlue representative replied to the tweet, recommending that the customer clear their browser history and try again.
The assumption, the lawsuit alleged, was that JetBlue was using customers’ personal data — anything from income to recent search history — to set prices, based on factors like their wealth, frugal or profligate spending habits, signs they might have a desperate need for the ticket. The practice is commonly called surveillance pricing or personalized pricing. JetBlue denied that they used it.
Lawmakers and regulators, both at the state and federal levels, have taken note. Amid a growing concern over how corporations set their prices, California lawmakers are debating a bill that would ban surveillance pricing.
But while some pricing experts are cheering on these bans, others think lawmakers are too quick to restrict a practice they say could actually lead to lower prices for consumers.
The backers of Assembly Bill 2564 were enthusiastic about its chances for passage, though the legislative year will come to an end Monday at midnight. The Assembly approved a version of the bill in May, and the state Senate was set to consider it Sunday. If passed, as backers hoped, it would return to the Assembly with still a full day for another vote.
A practice as old as time
Personalized pricing has been going on for centuries. For thousands of years, a seller at a market might have used visual and social cues — what a customer was wearing, what the seller knew about their family — to quote a price.
“The seller sized you up and thought you were wearing a really nice pair of sandals and decided to charge you more,” said Lindsay Owens, the author of a forthcoming book on pricing and the executive director of the Groundwork Collaborative, a nonprofit that conducts research on corporate pricing.
But one price for all, a concept conceived by the Quakers, who believed that charging different prices to different people was morally abject, became commonplace.
And now with consumers accustomed to the fixed price tag, companies are returning to differential pricing. With the surveillance model, a retailer or a third-party company uses personal data to pinpoint what economists call a customer’s willingness to pay. How extensively you have searched a product might indicate your eagerness to buy. Your income, your ZIP Code, your family situation might suggest how much you can afford.
“You could make an analogy to personalized advertising. How we’re constantly getting ads on Facebook that are really tailored to us,” said Abbey Stemler, a professor of business law and ethics at Indiana University’s Kelley School of Business.
“They’re using that same type of technology where they’re taking our individual data acquired by the companies themselves, or purchased through data brokers.”
But, Stemler said, there isn’t much empirical data to show how pervasive this practice is, mostly just anecdotes.
In 2022, the state of California reached a $5-million settlement with retail giant Target over allegations that the company was charging different prices for the same product on its mobile app depending on a customer’s location.
An investigation from Consumer Reports published last year detailed how grocery retailer Kroger built demographic profiles of shoppers, which it used to create personalized discounts. In one example, much of the information gathered about a customer was wrong, including his level of education and how many kids he had.
“It’s potentially extremely unfair, because sometimes those judgments aren’t correct,” said Jen King, a privacy and data policy fellow at the Stanford Institute for Human-Centered Artificial Intelligence.
In 2024, the Federal Trade Commission surveyed eight businesses that had advertised their use of AI and real-time data to determine prices, including Mastercard, JPMorgan Chase and retail pricing software companies.
The federal agency found that companies collected data on consumers’ IP addresses and language settings, but also much more granular information; one respondent said they could gather data on whether someone highlighted the name of a product on a webpage or how far down the page someone scrolled while shopping.
All of this could be used to infer how price-sensitive a customer was, or how willing they were to pay more for a product. In one use case, FTC researchers wrote that if a customer is profiled as a new parent and is shopping for a baby thermometer, that customer might be shown higher-priced versions of the product first.
If that same parent chooses a faster shipping option, a company might infer desperation for an item and conclude this person could be charged more in the future.
Does surveillance pricing lead to higher prices?
Assembly Bill 2564 would make it illegal for retailers to use surveillance pricing, defined by lawmakers as the use of personal data to set a price using any form of technology, including AI.
Other states have passed laws restricting the practice, including New Jersey. In New York, a bill banning surveillance pricing is awaiting the governor’s signature. Earlier this month, the FTC signaled its intention to go after companies that don’t disclose their use of personalized pricing.
Assemblymember Christopher M. Ward (D-San Diego), author of the California bill, said it addresses a recognized “unfairness” among customers.
“I think that there’s a fairness question,” Ward said. “That because you live in California and we have, on average, higher median incomes, some computer algorithm would infer that you are able to pay higher prices than somebody who lives in Arizona.”
Polls show that 3 in 4 Americans worry about how companies and the government are using their personal data. And in a poll from the Groundwork Collective, 76% of respondents said they thought it was unfair for retailers to charge different prices for items based on a customer’s personal data.
But what few studies exist on surveillance pricing have shown that while it enables companies to earn more, that doesn’t always translate to higher costs for consumers.
In a 2022 study, researchers looked at the prices ZipRecruiter, a job posting site, charged employers wanting to access their services.
When ZipRecruiter used data about these businesses to set personalized prices, researchers found that more than 60% of customers got a price lower than an optimal rate researchers had calculated. The company also made more money with personalized pricing because it was able to interest more buyers.
“When we start assuming personalized pricing might be bad,” said Jean-Pierre Dubé, a professor of marketing at the University of Chicago Booth School of Business, “we might be disadvantaging customers who may be benefiting.”
Some of those pushing back against California’s surveillance pricing bill argue this point — that it could make it harder for companies to offer lower prices — although the current version of the bill allows for certain discounts, including when a customer is about to end a subscription. Colorado Gov. Jared Polis vetoed a similar bill in June, saying the language was too broad and would, in his opinion, restrict companies from lowering prices, not just raising them.
Even if this is so, experts say personalized pricing could lead to a diluted understanding of what one item should cost.
“The lack of transparency can just end up unmooring you,” King said. “You’ll always be guessing, trying to figure out what is this thing actually, what is it, what should it cost? Am I actually getting a discount? Am I getting a deal?”
This article is part of The Times’ equity reporting initiative, funded by the James Irvine Foundation, exploring the challenges facing low-income workers and the efforts being made to address California’s economic divide.