According to PYMNTS.com, the Federal Trade Commission (FTC) has sent Instacart a civil investigative demand seeking information about its Eversight AI pricing tool, as reported by Reuters on Wednesday, December 17. This probe follows a recent Consumer Reports study, published last week with advocacy groups Groundwork Collaborative and More Perfect Union, which involved 437 shoppers across four cities. The study found that different shoppers saw vastly different prices for the same products from the same stores on Instacart, with an average 7% difference in total cost for identical grocery lists. Alarmingly, some shoppers encountered prices up to 23% higher than others for the exact same items at the exact same time. The FTC stated it is “disturbed” by the alleged pricing practices, though it noted a probe launch does not prove wrongdoing. Instacart, in a blog post last week, argued the study inaccurately conflated “A/B price tests” with “dynamic pricing” and denied using personal user data to set prices.
The FTC probe in context
Now, here’s the thing. The FTC isn’t messing around with “dark patterns” and algorithmic pricing these days. They’ve been on a tear, looking at everything from subscription traps to deceptive design. So an AI tool that lets retailers run massive, automated price experiments on live customers? That’s basically a giant red flag waving in their faces. Instacart’s defense hinges on a technicality: they say it’s not “dynamic pricing” (which changes based on demand) but “A/B testing” (which is randomized). But to the shopper who just paid $12 for a box of cereal that someone else got for $10, that distinction feels meaningless. The outcome is the same: you paid more for no clear reason.
Broader market impact
This is a huge deal for the entire grocery delivery and e-commerce space. Instacart is trying to position itself as a tech platform that provides tools to its retailer partners, like Kroger or Albertsons. But when those tools create consumer distrust, the platform itself takes the heat. And in a market where profitability is still elusive for many delivery services, the temptation to use AI to squeeze out every last cent of margin is intense. The risk, though, is a massive consumer backlash. The PYMNTS data cited is key here: with 42% of consumers living paycheck to paycheck out of necessity, even a perceived 7% unfair price hike feels like a betrayal. If shoppers start believing the app is rigged against them, they’ll just go back to the store. So this FTC probe isn’t just about one tool. It’s a test case for how much opacity in algorithmic pricing the market—and regulators—will tolerate.
What happens next?
Instacart is in a tough spot. They have to cooperate with the FTC, which will mean handing over a ton of data on how Eversight actually works. Their claim that they don’t use personal or behavioral data will be put under a microscope. Can they prove the price tests are truly random and not subtly influenced by a user’s order history, location, or device? I’m skeptical. Even if they’re technically clean, the perception is already damaged. The company will likely have to make the tool far more transparent or give users a way to opt out of pricing experiments. For the rest of the tech world, it’s a stark warning. As AI gets better at micro-targeting, regulators are getting better at spotting potentially unfair practices. Building your business model on secretive A/B testing might just be the next regulatory frontier.
