WASHINGTON — The Federal Trade Commission proposed on Monday that large businesses in several AI-negotiated consumer markets be required to show customers a standard offer before using personal profiles or purchasing agents to determine individualized terms.

The proposal, the first major federal action associated with the consumer campaign known as the Right to a Human Price, would cover large hotel groups, airlines, rental-car networks and telecommunications providers. It would not apply to groceries, ordinary mass retail or other markets where posted prices remain the norm.

Under the FTC's proposed Anonymous Reference Offer Rule, a covered seller would have to make one standard package available in both human-readable and machine-readable form. The offer would have to appear before the seller received behavioral information about the customer or began a negotiation with the customer's autonomous agent.

The reference offer could still vary with date, location, capacity and market conditions. It would not have to be the lowest available price, and businesses could continue offering discounts, loyalty benefits and negotiated bundles.

“A customer may still delegate the entire transaction if that is what they prefer,” FTC Chair Lillian Vey said in announcing the proposal. “What cannot disappear is an offer that exists before the seller knows who is asking.”

Consider a hotel room. A public reference offer might include a standard room, cancellation until the evening before arrival and an 11 a.m. checkout for $342. A purchasing agent could ask the hotel's system for a cheaper non-refundable stay without housekeeping, or pay more for a 3 p.m. checkout, airport transport and flexible cancellation.

Those customers are not necessarily being charged different prices for the same product. They have bought different combinations of time, risk and service. The rule would preserve that negotiation while giving both customers a common point from which to evaluate it.

The sharper concern arises when two people request substantially the same package but receive different offers because a seller has inferred that one is wealthier, more loyal, in a hurry or less willing to leave. That practice, often called surveillance or willingness-to-pay pricing, is distinct from configuring a product around a buyer's stated preferences.

The proposal would restrict when sellers can use that behavioral information, but it would not require a consumer's agent to disclose everything it knows. Modern buyer agents routinely use private information — a traveler's need to stay near a hospital, for example — without transmitting the underlying reason or revealing how much the person might be willing to pay.

Consumer groups have condensed the rule into a simpler demand: “Show me the price before you learn who I am.”

That demand would have been familiar but largely unnecessary in most consumer markets two decades ago. In the late 2020s, regulators focused on dynamic pricing, behavioral profiling and early attempts to infer a shopper's maximum price. Most transactions still began with a figure displayed on a shelf, screen or booking page.

Purchasing agents improved rapidly during the 2030s. Travel, insurance, telecom and large service platforms opened machine-readable interfaces through which consumer agents could request offers from competing seller systems. People increasingly supplied goals and constraints rather than selecting a fixed product themselves.

Seller agents responded with packages tailored to those requests. A single hotel stay could become a private bundle of cancellation rights, housekeeping, loyalty credits, transport and departure time. By the late 2030s, some booking pages had reduced the posted price to a secondary option or replaced it with an invitation to ask an agent for current terms.

The change often worked as intended. Agents could search more sellers, negotiate details people ignored and force businesses to compete for a specific customer. But it also made a basic question harder to answer: What does this cost before either side knows anything about me?

Critics began comparing the experience to an American dentist's office in the 2020s — not because dentists used autonomous negotiation, but because a patient could know what service was needed without finding one clear and comparable price in advance. By the early 2040s, the comparison had spread from consumer groups into congressional hearings and state campaigns.

A public benchmark has uses beyond an individual purchase. Consumers can test whether their agents actually secured a better deal. Regulators can compare similarly situated buyers. Economists and journalists can track price changes in markets where a growing share of transactions otherwise consists of private machine-to-machine contracts.

Industry groups said the proposal could make those comparisons look more precise than they are. The Association for Negotiated Commerce said a mandatory reference package would combine terms that many customers did not want and could anchor negotiations around a regulatory benchmark, reducing the flexibility that allows agents to find lower prices.

“There is no single real price for a service whose timing, risk and features are being assembled for each buyer,” the group said in a statement. It said consumers voluntarily use purchasing agents because the systems save time and, across several covered markets, generally produce more suitable offers.

The FTC acknowledged that risk. Its proposal asks whether sellers should be allowed to publish several standard reference packages and how regulators should compare them with customized offers. It also seeks comment on whether a reference offer could become a high-priced default for customers who decline automated negotiation.

That question carries a class dimension. Premium purchasing agents can search more markets, understand more complicated contracts and combine travel, insurance and loyalty programs more effectively than basic agents. Wealthier consumers may therefore receive better terms not only because sellers favor them, but because they arrive with more capable representation.

Industry opponents call the Human Price a potential “sucker price” — a benchmark disproportionately paid by people who do not use sophisticated agents. Supporters respond that without a benchmark, those consumers would have even less ability to identify the disadvantage.

The FTC will accept public comments for 90 days before deciding whether to issue a final rule. Whatever form it takes, the proceeding treats visible prices less as a retail convention than as a piece of market infrastructure: information needed for competition, measurement and enforcement.

“A market can become extraordinarily good at calculating the optimal price for every person,” said Tessa Mirren, director of the Consumer Reference Project. “The question is whether it still owes society a price that belongs to no person in particular.”

Reader forecast

Do you buy this future?

When do you think U.S. consumer law requires a public reference price in some AI-negotiated markets?

Relative to this report's 2044 date. One prediction per browser. Predictions cannot be changed after submission.