Glen Hauenstein, the president of Delta, told investors in July that the airline would expand its artificial intelligence (AI) pilot programme to optimise costs for individual customers. This has led to greater attention to the possibility of sellers setting custom prices based on each customer’s personal information.
Delta has since maintained that AI is only used to enhance its long-standing dynamic pricing system, which allows it to adjust fares to generally balance supply and demand, rather than setting individual fares. It is difficult to understand how AI could be applied to improve those traditional market pricing metrics to the extent that Hauenstein suggested.
Concerns about AI-driven pricing models that favour sellers over consumers continue to exist. To address the massive information gap between buyers and sellers that facilitates this invasive practice, it is essential to confront the lack of consumer privacy protections in United States law. Additionally, consumers must be empowered to defend themselves.
Why bespoke pricing is problematic
With advancements in digital technology and unrestricted collection of customer personal data, sellers can now exploit sophisticated algorithms to violate customer privacy and profit from the outcome. Each customer can have a personal profile that sellers can use to determine whether they are willing and able to pay more than what the traditional supply and demand curves would indicate.
For over a century, sellers have posted their products’ list prices. These prices are uniformly available to all customers, who are anonymous to the seller at the time of posting. To assess demand and adjust the price accordingly, sellers have used market research.
Economists call this ‘consumer surplus’ — the benefit customers gain when they would have been willing to pay more than the list price. The seller reverses the customer’s role by offering custom prices. Individual consumers will no longer be able to see or compare the prices that corporations offer.
However, some customers are not offered lower prices by sellers out of altruism. Sellers want to increase sales and reduce excess inventory to boost overall profit. Historically, they have achieved this through sales. But when the current price is available to all, customers can plan their purchases, look for discounts, and compare prices.
With this strategy, customers who would be willing to pay more but wait for or find a lower price can take advantage of the consumer surplus. By eliminating this option, bespoke pricing allows the seller to keep the excess money from customers.
For decades, airlines have used their market knowledge to develop distinct seat categories at varying prices based on factors like class, time of year, and time of purchase. They expect demand to rise during the holidays or during significant events, and assume that families and individuals are likely to make the majority of their purchases well in advance.
AI-powered ‘invisible hand’
Of course, customised pricing isn’t always unfair. Completed sales in medieval bazaars came from negotiations, where a buyer could evaluate a seller and vice versa. This is still how many large purchases, such as homes and cars, are made.
However, for a modern industrialised economy, individual negotiations generally proved ineffective. List prices, set before sellers knew who would be purchasing, were posted and advertised starting in the 1880s. For consumers, the price tag provided anonymity.
The advent of online commerce gave consumers hope that it would increase their purchasing power by facilitating the comparison of prices and finding desired goods and services. This was true, and it did just that. However, the rise of custom pricing threatens that optimism. A new market for consumer data was created by online commerce, and it has since been gathered, organised, shared, sold, and cross-referenced on a massive scale.
How can consumers avoid giving up control of their personal information? One option is to persuade the government to restrict the use of custom pricing, which the Federal Trade Commission explored in 2024. Congress and state legislatures are currently considering legislation at various stages.
A crucial first step is enacting stricter data privacy legislation. In the absence of legal restrictions, consumers may also choose to address the issue on their own by hiding their identities and limiting seller profiling. For instance, they might use a virtual private network (VPN) to shop online. It will be difficult to return some of the power consumers once had due to bespoke pricing.
The industry’s vastly superior financial, technological, and lobbying resources hinder consumers and their advocates. However, for the online marketplace to continue to serve consumers, addressing bespoke pricing in some form is essential.
In addition to individual actions like using VPNs and legislative changes, raising public awareness and promoting digital literacy are equally important in addressing the challenges of custom pricing. Customers often don’t realise that every click, search, and cart abandonment leaves a trail of data that algorithms use to predict and control their purchasing habits.
This lack of transparency exacerbates the power imbalance between buyers and sellers. Public education about the collection, analysis, and monetisation of personal data is crucial. People can only demand greater accountability from businesses profiting from their data and make informed decisions about the tools and platforms they use if they have this knowledge.
Additionally, governments, civil society, and the private sector must work together to create ethical AI frameworks. In automated pricing models, these should prioritise non-discrimination, fairness, and transparency.
The same stringent regulations that apply to food and drug companies because of their potential effects on human health should also apply to businesses using AI in commerce to safeguard economic well-being. Independent audits of algorithmic pricing tools, transparency reports, and unambiguous opt-out procedures should be standard practices, not optional features.
Although the concept of “algorithmic fairness” is gaining traction worldwide, it is still not widely applied in practice. To enforce compliance in this rapidly evolving field, regulatory bodies must have both technical expertise and authority.
Furthermore, businesses that use data ethically and implement fair pricing practices should be rewarded, whether through financial incentives, public recognition, or trust ratings. In the digital economy, fairness will be redefined to benefit consumers, turning it into a competitive advantage rather than an afterthought.
If AI-enabled custom pricing is allowed to continue, it could gradually undermine decades of efforts to create consumer protections. Now is the time to act before custom pricing becomes so deeply ingrained in online sales that its effects are nearly impossible to reverse. Fair pricing should be protected for everyone in the online marketplace, not just the wealthy and powerful.
Experts from Northeastern University in Massachusetts have raised alarms about the risks to consumers. According to Kate Ashley, an associate teaching professor of supply chain and information management at Northeastern, AI flight pricing could reduce transparency, enable discrimination, and give companies too much power over what individuals pay, potentially exploiting travellers without their knowledge.
This observation comes as Delta Air Lines faces scrutiny after revealing plans to use AI to set ticket prices. Critics are questioning whether such technology could lead to personalised pricing that targets consumers based on their data.
Along with Ashley and her Northeastern colleagues, lawmakers have raised concerns as well. Democratic Senators Ruben Gallego, Mark Warner, and Richard Blumenthal asked whether Delta was using AI to set prices based on personal data. If so, they warned, it could lead to fare increases tailored to each customer’s personal “pain point.”
Christo Wilson, a Northeastern professor of computer science and founding member of the university’s Cybersecurity and Privacy Institute, made it clear with these words: “It very much sounded like this was being used for personalised, individual pricing. This is the kind of thing that gets people very upset. It doesn’t sound good.”
AI-driven pricing might make sense for businesses, but it risks taking advantage of consumers who don’t even know it’s happening. When companies use personal data to decide what each person should pay, fairness disappears. Ultimately, Prices should reflect value, not how much someone can be squeezed for. After all, we need rules that protect everyone, not just profits.