Instacart customers may be surprised to discover they are unwittingly paying more for the same items sold by some of America's major retail chains than their fellow shoppers.

A months-long investigation by the nonprofit organizations Consumer Reports and Groundwork Collaborative found that identical grocery items on Instacart could differ in price by as much as 23% from one customer to the next.

That's due to the platform's algorithmic pricing experiments, which place different price tags on identical products without revealing the discrepancies directly to shoppers, the report found. The AI model, which Instacart started implementing in 2022, sets grocery prices at some large retail chains that partner with the San Francisco-based delivery company, the groups added.

Variations in pricing for identical products are particularly acute in online shopping, given that customers don't have the same reference points as in physical stores, said Neil Saunders, managing director and analyst at GlobalData.

"You sit in front of your phone or your browser, you're shown the price, and you don't know what everyone else was shown," he said.

Justin Brookman, director of digital marketplace policy for Consumer Reports, agreed. "Traditionally, we haven't had to worry about this sort of thing. We would go to the supermarket and pay what was on the shelf," he told CBS News. "Now, I think people are going to be worried: Am I getting ripped off?"

Report finds major price discrepancies

Consumer Reports and Groundwork Collaborative based their findings on data gathered from online shopping sessions, in which hundreds of volunteers shopped on Instacart for identical baskets of goods from Safeway and Target.

Of the 437 participants, every single one was exposed to algorithmic price experiments, according to the report. The investigation also found evidence of price experimentation at Albertsons, Costco, Kroger and Sprouts Farmers Market.

According to Consumer Reports and Groundwork Collaborative, Instacart has framed price differences between customers as "negligible." However, the scope of the price experiments is "far broader and more costly to some consumers than has been publicly acknowledged," the organizations said in the report.

The investigation found that some products had as many as five different price points, with variations ranging from as little as 7 cents to $2.56 per item. During one test conducted for a Safeway in Seattle, the price for a box of Wheat Thins differed by as much as 23%. 

All told, the price variations could cost families $1,200 a year, based on how much Instacart says the typical household of four spends on groceries, the report found.

Some products were unscathed by the price experiments. For example, at the Seattle Safeway, volunteers did not observe any price variations for items such as Premium brand saltine crackers, Heinz ketchup and Barilla farfalle pasta.

Safeway didn't immediately respond to a request for comment.

Instacart told CBS News in a statement that 10 of its retailer partners are using pricing experiments, but did not specify which ones.

"Just as retailers have long tested prices in their physical stores to better understand consumer preferences, a subset of only 10 retail partners — ones that already apply markups — do the same online via Instacart," a spokesperson said in an email. "These limited, short-term and randomized tests help retail partners learn what matters most to consumers and how to keep essential items affordable."

Customers were also shown different "original" prices, making some savings appear larger, the report found, a concept known as "fictitious pricing." Amazon was sued this October for allegedly using this tactic during its summer Prime Day sale.

The rise of AI pricing models

Dynamic pricing, which sets consumer prices based on demand, has long been part of the e-commerce landscape. However, while it's a common tactic in certain industries like travel, where supply is fixed, consumers may be surprised to see it cropping up in the grocery space, Saunders said.

"In retail, especially grocery retail, people expect the price to be the price," he said. "They don't expect [it] to yo-yo up and down."

Another issue, as the report points out, is that consumers often have no clue that they may be paying more for the same item. The Federal Trade Commission (FTC) Act prevents "unfair or deceptive acts or practices," including "misleading cost or price claims." However, Saunders said he doesn't believe companies are legally obligated to tell customers that they are part of pricing experiments.

"Instacart shoppers we spoke to say they were unaware that they were participants in active Instacart pricing experiments and view the practice as manipulative and unfair," the report says.

One example of dynamic pricing is surveillance pricing, a tactic employed by companies that involves using a shopper's behavior and personal data to make price evaluations, according to the FTC. Saunders said surveillance pricing has been supercharged by artificial intelligence, which can assess vast amounts of data at rapid speeds to determine price levels.

Instacart told CBS News that it does not use personal, demographic or user-level behavioral data to set prices.

As pricing models like Instacart's become more sophisticated, Brookman said he's concerned about what it could mean for consumers.

"I'm worried companies are going to use these sorts of algorithmic pricing models to charge us the maximum amount we're willing to pay, which is good for their profits, but not so good for [our wallets]," he said.