Walmart Patents Reveal Algorithm That Could Raise Prices Based on Shopper Profiles
Baca dalam 60 detik
- Groundwork Collaborative menemukan belasan paten Walmart yang mendukung eksperimen harga, penargetan kupon, dan tampilan harga berbeda per pengguna.
- Walmart membantah menerapkan harga personal, tetapi angka internal menunjukkan pengguna asisten AI Sparky memiliki nilai transaksi 40% lebih tinggi.
- Ketiadaan regulasi transparansi harga otomatis membuka celah antara kemampuan teknologi dan praktik yang benar-benar dijalankan.

A recent report from the watchdog group Groundwork Collaborative reveals that Walmart holds a number of patents that allow algorithms to adjust prices and promotions based on each shopper's profile. The findings raise a fundamental question: is the price shown in the Walmart app to one person the same as what another person sees?
Groundwork identified more than a dozen patents that, by their characterization, go beyond the logic of ordinary loyalty programs. One patented concept describes gradual price experiments—raising prices little by little to gauge how deep a consumer's pockets are. Another patent designs a system that withholds coupons from customers the algorithm predicts will still pay full price. Yet another patent allows different prices to appear in search results; Groundwork offers a hypothetical example of one shopper seeing $90 while another sees $110, based on shopping history, location, and other identity data.
These patents are not in themselves proof of an active practice. Groundwork itself stresses that the $90 versus $110 scenario is their interpretation of patent documents, not a price ever shown to a real customer. However, the report also highlights a patent that can estimate a child's age from purchase history and adjust product recommendations as the child grows, as well as an electronic shelf label patent granted in 2023. Several other applications cover facial recognition, Wi-Fi tracking, and physiological data such as heart rate—the same categories of data that surveillance apps have demonstrated for profiling targets.
"Corporations don't spend years building sophisticated pricing tools, hiring experts, and filing patents just to let them gather dust," said Groundwork Collaborative CEO Lindsay Owens.
Walmart responded with a firm statement. EVP Dan Bartlett asserted that the company does not and will not use personal information, income, shopping history, urgency, or a customer's willingness to pay to set individual prices. Their slogan: "We set prices on products, not people." Walmart also argues that patents are merely descriptions of possible inventions, not active business practices, and stated it will not use or renew patents governing cart-based price changes through Scan & Go and electronic shelf labels.
The 40% figure from Walmart's own investor communications only sharpens the tension. Groundwork interprets that jump in transaction value as evidence that AI shapes shopping behavior for Walmart's benefit. Walmart rejects that "upcharging" framing. A higher cart value does not automatically prove personalized pricing—it could stem from more items, product substitution, or broader recommendations. Still, the question remains valid: what exactly is Sparky optimizing, and should consumers know the answer before using it?
For Indonesian consumers, this issue is not merely a story from a distant land. Domestic retail giants and e-commerce platforms are beginning to adopt AI for price personalization, promotions, and recommendations. Without an adequate transparency framework, the gap between "can" and "will" can easily become an unmonitored practice. Regulators in Indonesia—from KPPU to Kominfo—have no specific rule requiring disclosure of the logic behind automated pricing. Yet citizens' behavioral data is increasingly easy to collect through apps, loyalty programs, and digital wallets.
The core problem is not the patents, but the absence of transparency that leaves the distance between capability and restraint unchecked. A patent portfolio shows what a company can build; those documents do not guarantee what they will choose not to do. Shoppers who use the app, Scan & Go, or the Walmart site deserve to know what data Sparky accesses, whether promotions are equally available to all customers, and whether price recommendations vary by individual profile. That demand is not excessive—it is basic disclosure that no regulator has yet required.
The right response to this report is not panic, but pressure: on Walmart, on the retail industry broadly, and on supervisory authorities. Transparency over automated pricing decisions should be mandatory before the line between capability and practice blurs. The question now: how fast will regulators—in the United States and Indonesia alike—move before algorithms set prices that can no longer be accounted for?



