Yale School of Management: surveillance pricing is just the beginning. AI agents will be the real test of corporate trust

For all of 2025, despite a flurry of proposals, not a single state managed to ban “surveillance pricing.” This spring, that changed. In April, Maryland became the first state to prohibit food retailers and delivery services from using consumers’ personal data to set prices. In June, Connecticut became the second. California and New York are considering similar measures as part of a broader wave of efforts to limit surveillance pricing. Yet the real issue is not fundamentally about pricing. It is about how companies choose to use data, algorithms, and increasingly AI: when technology lets them understand customers and workers in unprecedented detail, will they use that insight to create value or to extract it? The question is whether companies are pricing the transaction or the person. Consider two identical Uber requests from Midtown Manhattan to Newark Airport. Most riders understand why the trip costs more on a rainy Friday afternoon than on a quiet Sunday morning; adjusting prices for weather, traffic, or supply is a transparent way to balance a market. But consumers increasingly question something different: two riders standing on the same corner at the same moment paying different prices based on their data profiles, purchase histories, devices, or inferred willingness to pay. In the first case, the platform is pricing the ride. In the second, it is pricing the rider. The Federal Trade Commission made the stakes concrete in a 2025 study, showing how algorithms drawing on personal data can infer when consumers may have fewer alternatives, greater urgency, or a higher willingness to pay—and adjust prices or offers accordingly. The same logic operates on the other side of the marketplace. An algorithm may offer a lower payout to a driver it predicts will accept anyway—because she is nearing a daily earnings goal or unlikely to switch apps. When companies shift from pricing a transaction’s conditions to exploiting the vulnerabilities of the people in it, they drift