I remain concerned that AI will make online shopping more chaotic and untrustworthy.
But my impression has shifted from ABSOLUTELY NO to hmm … maybe this is a decent use of text-summarizing AI — as long as you learn to read Amazon’s AI digests with a savvy eye.
Amazon founder Jeff Bezos owns The Washington Post. I’m writing about these AI summaries because Amazon’s customer reviews have already influenced the entire system of online feedback for hotels, restaurants, movies and doctors.
Like it or not, Amazon’s experiment with AI is a harbinger of a future in which we’ll need skills to sift through AI-written text and AI-generated images.
Take those AI summaries with a grain of salt
I picked at random an Amazon product listing for this Stanley insulated mug. Scroll way down the page to the section for customer reviews.
One of the first things you’ll see is a few sentences of what “Customers say” about the product. Amazon says that is a distillation by AI from the text of nearly 40,000 customer reviews.
(It appears that you must be logged in to your Amazon account to see AI summaries. Not all products have AI-generated summaries of customer feedback.)
Juozas Kaziukenas, founder of the e-commerce research firm Marketplace Pulse, pointed out that since Amazon started the AI-generated review summaries last year, the company has tweaked them to highlight terms or features that apparently come up a lot in customer ratings.
The positive features are highlighted in green and the negative or neutral feedback is in yellow and gray. You can see that in the Stanley mug AI summary, with “Scratching” and “Value” in those undesirable colors.
If you like to get a gist of what shoppers thought of a product, Amazon’s AI summary can spare you from skimming the reviews yourself.
When my colleague Danielle Abril was hunting recently for pet hair clippers, one AI blurb mentioned fur getting caught in a product’s guard and the blades not being sharp. That was enough for her to rule it out as a possible purchase.
But as with Amazon reviews in general, the AI summaries might be incomplete or untrustworthy.
I looked up a (human) hair removal product that some of the reviewers complained had irritated their skin or didn’t work as promised. Those complaints weren’t mentioned in the AI summary of the product reviews.
The AI summary also describes the product as a “foot scraper.” It’s not.
Bloomberg News recently looked at dozens of AI review summaries and found in some cases they underplayed customers’ negative feedback and exaggerated them for other products.
And, of course, if the reviews themselves are misinformed or rigged, a summary of junk customer feedback will also be junk.
Amazon said the company is “seeing positive feedback on our review highlights from both customers and sellers” but that it will “continually improve the review highlights experience over time.”
What’s coming next for AI in shopping
Amazon, eBay and Shopify are also experimenting with using AI to spit out descriptions of products from a photo or a few keywords.
Some of this AI-generated text will be better than the confusing product listings you sometimes read online. A lot of it will be worse.
A bunch of technology companies, including Amazon and Meta, are also betting that AI will be better and cheaper than current methods for creating product advertisements to clog your online shopping results and social media feeds. Hooray, right?!
Kaziukenas is worried, too, about the potential for misuse of AI in online reviews. You could imagine companies trying to pay off or nudge reviewers to repeatedly mention keywords in the hopes that AI will include them in a review summary.
“If this grows, that will become yet another part of the Amazon game — getting the right kind of reviews mentioning the right kind of things to make the algorithm summarize it best,” Kaziukenas said.
This has already happened with Google’s web search AI.
Websites stuff themselves with frequently Googled words so you might find them when you search for “easy chicken dinners” or “Mother’s Day gifts.” (The Washington Post does this, too.)
Needing to use magic words for web search robots is why recipe websites drown you with pages of text before you get to the recipe. It results in gibberish product listings like this one: “Reusable Stainless Steel Water Bottle Travel Mug Cupholder Friendly Insulated Cup Holiday Gifts for Women Men Him Her (Lavender).”
In other words, the web is already written for AI in ways that make it unreadable for humans. I worry that even more AI in shopping and beyond will make the problem much worse.
