Artificial intelligence is giving airlines a tighter grip on ticket pricing, and Bloomberg reports that the change is making cheap flights harder to find.
For years, airlines relied on analysts to write pricing rules by hand. A flight that reached one-quarter occupancy, for example, might trigger a 20% fare increase. Those systems were built by people, and the gaps between rules created openings where travelers could sometimes pick up unusually low fares.
Bloomberg said AI is now closing those gaps. Pricing systems can weigh dozens of variables in real time and optimize fares across an airline’s network at the same time, helping carriers capture more revenue. On popular routes, that means accidental low-price seats are likely to become less common.
Delta and Fetcherr have become the center of the dispute
The issue has been debated in the U.S. for about a year, with Delta Air Lines’ work with Israeli startup Fetcherr acting as the trigger.
According to the report, Fetcherr uses generative AI to replace human analysts in the pricing process, running analysis and generating quotes around the clock. Its client list also includes Virgin Atlantic, WestJet, Azul and Aerobus.
Delta said in January 2025 that about 3% of its domestic fares were then being priced by AI, and that it aimed to raise that share to 20% by the end of the year. After that disclosure, Democratic senators sent a letter questioning whether personalized pricing could push fares up to each consumer’s maximum pain point.
The Federal Trade Commission had already started examining surveillance pricing in July 2024. In January 2025, it released preliminary findings saying companies were using algorithms to automatically set prices based on personal data, a practice that could lead to higher prices for vulnerable groups predicted to spend more, including people who had recently given birth and needed to buy infant formula.
Delta’s response was direct: it said it does not provide any personal information to Fetcherr, and that fares have never been based on personal data.
The loss of cheap tickets does not necessarily depend on personal data
The report says this is the part most likely to be misunderstood.
Cheap fares are disappearing not only because AI might identify who a traveler is or what that person may be willing to pay, but because AI can understand the state of a flight better than older human-set pricing rules. A rule such as raising prices by 20% once a plane is 25% full left room for pricing errors because it was coarse.
A system that can process remaining seats, competitor changes, historical curves, tourism and holiday events, weather, fuel prices and exchange rates in real time does not need a passenger’s name, age or income to push fares close to the upper limit the market will bear at a given moment.
Under that view, the privacy question and the fare question are separate. Regulators have largely focused on the first one, while the second may be what actually causes consumers to pay more.
Bryan Terry, managing director at Alton Aviation Consultancy, put it plainly: consumers should expect airlines to become smarter about pricing, raising fares when there is room and cutting them only when they need to stimulate demand.
A structure that resembles MEV in crypto markets
The report draws a comparison with maximal extractable value, or MEV, in crypto.
In blockchain markets, a user’s transaction first enters the public mempool, where others can see it before it is confirmed. Bots can then move ahead of that order and shift the price, leaving the user with a worse execution. That is the basic structure of MEV.
Bloomberg’s framing is that airlines are doing something similar to passengers: intent is visible before execution, and the counterparty adjusts pricing before the purchase is completed, producing a less favorable outcome for the user.
The difference is in how the industries respond. Crypto has spent years trying to limit MEV through private order flow, batch auctions and encrypted mempools, with privacy-focused projects in particular trying to prevent intent from leaking early. One draft regulatory proposal in the European Union even described MEV as “illegal market abuse” and called on exchanges to report suspicious transactions.
In aviation, the same broad logic is being presented as an upgrade in revenue management.
Regulatory activity in the U.S. is already underway
The input states that more than 40 bills across 24 U.S. states are addressing surveillance pricing, while the FTC has opened a civil investigation into the airline industry.
For now, much of the debate still centers on whether airlines are using personal data in pricing. Bloomberg’s core point is different: even without personal data, AI can reduce the chances of finding a bargain by filling in the cracks left by human pricing rules.