How AI-driven pricing systems determine what you’re willing to pay

They know who you are, where you live, how much money you make and where you spent your last vacation.

They’re watching what websites you visit, tracking your mouse movements while you’re there and what you’ve left behind in virtual shopping carts. Mac or PC? iPhone or Android? Your preferences have been gathered and logged.

And they’ve got the toolkit, powered by artificial intelligence software, to assemble all this information to zero in on exactly how much you’re likely willing to pay for any product or service that might strike your fancy.

The “they” is a combination of retailers and service providers, social media operators, app developers, big data brokers and a host of other entities with whom you have voluntarily and involuntarily shared personal and behavioral information. And they’ve even come up with new labels to make you feel better about the systems that are using your personal data to set a custom price.

Dynamic pricing. Personalized pricing. Even “discount pricing.”

FTC investigates surveillance pricing

But the Federal Trade Commission and others have another name for it: surveillance pricing.

In an ongoing investigation launched last year, the FTC is looking into the practice of surveillance pricing, a term for systems that use personal consumer data to set individualized prices — meaning two people may be quoted different prices for the same product or service, based on what a company predicts they are willing — or able — to pay.

Part of the FTC’s mandate includes working to prevent fraudulent, deceptive and unfair business practices along with providing information to help consumers identify and avoid scams and fraud, according to the agency.

In a preliminary report in January, the agency highlighted actions it’s already taken to quell the rise of surveillance pricing amid its effort to gather more in-depth information on the practice:

• One complaint issued by the FTC included allegations that a mobile data broker was harvesting consumer information and sensitive location data, including visits to health clinics and places of worship which was later sold to third-parties.
• The agency said it issued the first-ever ban on the use and sale of sensitive location data by a data broker which allegedly sold consumer location data it collected from third-party apps and by purchasing location data from other data brokers and aggregators.
• Another FTC complaint alleged that the data broker InMarket used consumers’ location data to sort them into particularized audience segments — such as “parents of preschoolers,” “Christian church goers,” “wealthy and not healthy” — which it then provided to advertisers.

Last year, the FTC issued orders to eight companies that offer surveillance pricing products and services that incorporate data about consumers’ characteristics and behavior. The orders, according to the agency, seek information about the potential impact these practices have on privacy, competition and consumer protection.

“Firms that harvest Americans’ personal data can put people’s privacy at risk. Now firms could be exploiting this vast trove of personal information to charge people higher prices,” said then-FTC Chair Lina M. Khan. “Americans deserve to know whether businesses are using detailed consumer data to deploy surveillance pricing, and the FTC’s inquiry will shed light on this shadowy ecosystem of pricing middlemen.”

AI-driven consumer profiling

George Slover, general counsel and senior counsel for competition policy at the Center for Democracy and Technology, has his own term for the practice of using personal information to construct prices for individuals: “bespoke pricing.” He said it poses a fundamental threat to consumer welfare and free market principles.

Proponents of individualized pricing systems have argued that the method can send prices both directions — higher prices for some, lower prices for others. But Slover warned that, unlike uniform pricing, bespoke pricing enabled by “big data” and artificial intelligence gives companies little incentive to offer discounts to those who can’t afford market prices.

“Theoretically, maybe,” Slover told the Deseret News. “But as a practical matter what the sellers will do is maximize their prices. There’s a lot less incentive to lower the price for someone than raise the price for someone.”

Slover characterized the FTC investigation as “very useful” and said it could reveal more about the methodologies behind bespoke pricing and potentially lead to appropriate restrictions under existing law.

For now, he said, consumers have few defenses beyond masking their data.

“Potential ways for consumers to change their profile include working through a (virtual private network) internet connection, using an anonymized intermediary or even setting up a bogus, fictional profile … looking to reduce how much they have to pay,” Slover said.

But he cautioned that vast troves of consumer data on virtually every internet user have already been harvested and repackaged by brokers.

Slover, who has worked on antitrust and competition law for more than 35 years including stints with the U.S. Department of Justice and U.S. House of Representative’s Judiciary Committee, tied the debate over surveillance pricing to the broader need for comprehensive privacy protections.

“My organization, the Center for Democracy and Technology, was founded 30 years ago when the internet was getting off the ground,” he said. “One of the issues we’ve been focusing on since the beginning is the privacy of data … and getting Congress to implement a strong, comprehensive privacy law.”

Looking to protect data privacy

Utah state Rep. Tyler Clancy, R-Provo, said he’s not willing to wait for Congressional action on data privacy, which he sees as the critical underlying issue behind surveillance pricing.

“The privacy aspect is the biggest issue for me,” Clancy told the Deseret News. “Companies use that data to do business … but this is an area where we need some guardrails.

“If you’re creating a price for someone from immutable characteristics — race, faith, gender, ethnicity — that runs into constitutional concerns.”

Clancy said he is exploring “consent provisions” to ensure Utahns can know if their data is being used in pricing systems. He cited the response to recent news stories generated after a Delta Air Lines executive indicated in an earnings call that the carrier was testing out artificial intelligence technology that could set fares based on “the amount people are willing to pay for the premium products related to the base fares.” Delta issued a follow-up statement, clarifying that it was not using any manner of “individualized pricing” to set air fares based on customer data.

“There is no fare product Delta has ever used, is testing or plans to use that targets customers with individualized prices based on personal data,” said Delta chief external affairs officer Peter Carter.

But that qualification came after an uproar had already begun, Clancy said, and the issue has drawn interest from across the political spectrum.

“When this news story originally broke, it was a shock to people on the political right and the political left,” Clancy said.

Clancy said he’s working on a proposal for the 2026 session of the Utah Legislature and aims to compel transparency in how business entities use personal information in pricing systems for products and services.

“Sunshine is the best disinfectant and overall that’s the goal I’m trying to achieve here,” Clancy said. “It will lead to a better and freer market and that’s a win for everyone.”

Targeted advertising came first

When it comes to names, BYU marketing professor John Howell isn’t a fan of surveillance pricing, a term he believes is unnecessarily inflammatory. But he says the growing controversy over the practice isn’t about whether it’s possible or not, but whether consumers will tolerate it.

“This isn’t a new phenomenon,” Howell said in a Deseret News interview. “We’ve been paying attention to this at least from the early 2000s, individual level pricing, first-degree price discrimination. Charge every person a specific price based on their willingness to pay.”

Howell said economists have long predicted the advent of such models, but noted industry made a first stop in the advertising realm.

“It’s been conjectured as coming for at least 20 years,” he said. “Industry went to targeted advertising before targeted pricing. And I’m surprised that they went for that when targeted pricing is more profitable.”

Even if the practice is logical from a business perspective, Howell said consumer reaction is also predictably negative.

“Any time customers start to see price discounts, or individual pricing, they absolutely hate it,” he said.

And that tension isn’t new either. Howell highlighted that the Sherman Antitrust Act, passed in 1890, was “largely inspired by price discrimination by the railroad industry back at the turn of the 20th century.”

From an academic perspective, Howell said, price discrimination should benefit more consumers than it harms.

“The theory is, price discrimination almost always leads to lower average prices,” Howell said. “That doesn’t mean that every person is going to pay less. Also, it generally increases the availability of goods and services to lower-income customers.

“If you have price discrimination you can charge the people that can afford to pay a lot and less for those who cannot afford it. That’s generally what happens.”

In practice, however, Howell notes entities with the most data, and by default the most control over individual pricing systems, can easily disrupt the realm of competition.

“It’s not the theory of price discrimination that’s been disrupted,” he said. “What the data has done in our current economic system is tend to reward big players with increased power and lock out all the smaller players. As soon as they’re not competitive, the theory falls apart. If I was a policy maker, that’s what I would target, keeping big tech from accumulating so much power.”