How much did AI charge you for that plane ticket?

I am something of an airline points obsessive, or worse. I lurk on blogs and some more dodgy corners of the Internet to learn the latest strategies and promotions. I know the stakes are low, but it’s fun — even the outrage, the latest of which was ignited among the Delta flying community when the airline announced it was using AI to set prices.

Its mistake was using the term “dynamic pricing, ” which people interpret as personalized pricing, like charging people more depending on why they’re flying. Delta is not doing that (yet), at least not with AI. But it is coming, not only for airlines but also for many consumer businesses. And we will all hate it — but we should all prepare for it.

“Dynamic pricing,” like consumption taxes or market-rate-rent, is something only an economist can love. Dynamic pricing usually means a price that changes depending on when you buy — say, when Wendy’s wanted to charge less for burgers outside of peak lunch hour. The plan caused such an uproar that it was abandoned. But it would have been better for customers, who could have saved money by eating lunch later (or earlier, if they were hungry).


Dynamic pricing is not the same as personalized pricing, which is based on someone’s personal characteristics, such as how much they can pay or how much they value the good or service. People hate dynamic pricing, but they tend to hate personalized pricing even more.

Neither dynamic nor personal pricing is new. In some ways, they represent a return to a more traditional way of pricing, when people haggled in markets. Prices then often depended on the time of day, or a customer’s bargaining skills, or even the mood of the seller. In this era of mass consumption, however, something about this practice feels wrong and makes people angry — even in the economically savvy points community.

Dynamic pricing powered by technology became more noticeable with ride-sharing apps that charge higher prices when and where there is more demand. But it long predates Uber — movie theater tickets are cheaper during the day. Airlines also practiced dynamic pricing before the AI revolution, charging different prices for the same flight depending on when people buy their ticket, how full the plane is, the day they fly, and so on.

Airlines face a difficult problem setting their prices. Earning a profit requires them to fill planes — despite no-shows and weather delays. They serve customers with wildly different pricing sensitivities and travel needs. The market is also very competitive, with many people searching competitors’ fares (sometimes using AI!) to find the best price.

It is not surprising airlines are using new AI tools to make their prices even more dynamic — these complex problems are what AI is made for. And for fliers with more flexibility, it could mean cheaper air fares. And to be honest, personalized pricing already exists: People can pay more for an upgrade in class, a better seat, a checked bag or a meal. In all these cases, the airline is charging a price based on the customer’s ability to pay or comfort preference.

Dynamic and personal pricing is one reason that flying has become cheaper and more accessible over the years: Leisure travelers generally pay far less than business fliers. The difference now is that the algorithm will get better at setting personalized prices. It’s all but inevitable, in fact, that AI will be used to set personalized prices for pretty much everything sold online. If the data is there, sellers will want to use it.

Quite aside from its impact on consumers (see above: They’ll hate it), this new world has broader implications for the economy and society. How will it affect measures of inflation? Will more price-sensitive consumers get lower prices and a different inflation rate? What about data — will there be laws or norms about what companies can use? If someone spends months researching their dream vacation, bookmarking sites and messaging friends and family, and then finally decides to buy — will the seller know how excited they are, and jack up the price? What if their dishwasher breaks, and they’re desperate to buy? Will we all need to adjust the privacy settings on our browsers?

In a world powered by enormous amounts of data and the technology to make sense of it, more dynamic and personalized pricing is the future, whether we like it or not. When we buy online, maybe the prices we pay will be unique to us. It may be harder for retailers to pull off differentiated pricing in actual brick-and-mortar stores. So maybe in-person shopping is due for a comeback — though it’s unlikely in the travel industry, not least because we addicts need the web to feed our habit.