That lipstick review could have been written by AI

• Companies are using AI to help real people write product reviews.
• The beauty industry especially relies on product reviews to reach customers.
• But AI-powered reviews aren't necessarily good for customers trying to tell if a lipstick is any good.

I purchased the Haus Labs foundation after watching a glowing review of the product by a makeup influencer on TikTok.

Shopping online, it's sometimes hard to know how a foundation or lipstick shade will look on your skin — or how to know if that $60 serum really is going to exfoliate better than a $20 one.

And beauty products are highly subjective — what feels or looks good on one person might not on another.

That's where online reviews come in: For beauty, especially, reviews from influencers and regular people are crucial. And we're talking big business here. The beauty business is expected to become a $580 billion industry by 2027. Although several large conglomerates rule, it's still an industry where new brands can break out and become huge — Selena Gomez's Rare Beauty line is valued at $2 billion, for instance.

Good reviews — and lots of them — are also important for a product to stand out on a beauty site like Sephora, Ulta, or even Amazon. The reviews are so key that companies specialize in helping brands get good ones. Influenster is one platform where regular people can sign up to receive free products in exchange for writing and posting reviews.

Now, AI is in the mix.

Business of Fashion reports that AI-generated reviews for beauty products are making it harder for some shoppers to determine whether a review comes from a real person. For example, a company called Bazaarvoice has a new AI tool called "Content Coach" that says it helps consumers write beauty product reviews:

Bazaarvoice, a platform for user-generated content which owns Influenster and works with beauty brands including L'Oréal, Pacifica, Clarins and Sephora, has recently launched three new AI-powered features, including a tool called "Content Coach." The company developed the tool based on research showing that 68 percent of its community had trouble getting started when writing a review, according to Marissa Jones, Bazaarvoice senior vice president of product.

Content Coach gives users prompts of key topics to include in their review, based on common themes in other reviews. The prompts for a review of a Chanel eyeliner might include "pigmentation," "precision" and "ease of removal," for instance. As users type their review, the topic prompts light up as they are addressed, gamifying the process.

In this case, AI is helping someone who might not have the terminology to describe what makes a concealer feel so smooth on the skin. It could help a mediocre writer create a more coherent and knowledgeable-sounding review.

That's one of the things that generative AI is great for: taking a few bullet points of text and putting them into a few paragraphs with proper grammar and acceptable style. Amazon is reportedly testing out a similar tool to help people write reviews.

Bazaarvoice seems to be aware of some of the potential issues with AI-written content. "In terms of generative AI, if people can't trust whether a product review is genuine and written by a real user, they might choose to shop elsewhere," said Marissa Jones, the company's senior vice president of product. "So we as an industry need to think carefully about how we approach this technology in a thoughtful and ethical way."

Already, without AI in the picture, there have been some unfortunate and predictable problems with online reviews. Amazon has long struggled with fake reviews for its third-party sellers. Some beauty brands, like the skincare line Sunday Reilly, have gotten in trouble with the FTC for having their own employees write glowing reviews on the Sephora site. Fakespot, a company that detects fake reviews, estimated that 31% of reviews on the big retailers' sites — Walmart, eBay, Best Buy, Amazon — were suspicious in 2021.

When we typically look at reviews, we use our judgment to guess how trustworthy a reviewer is. If someone writes, "THIS FOUDNASHUN IS AWSUM U SHOUDL BUY IT IM 9 BTW" about the Haus Labs foundation I bought, you probably wouldn't take it as seriously as a longer, more eloquently written review that mentions the packaging, long-lastingness, texture, etc.

And some reviewers might be knowledgeable and good at assessing a lipstick, but bad at putting those skills into writing. For those cases, AI is great. But the fact that AI is suggesting which attributes to focus on in describing a product suggests that it's doing more than just helping clear writing.

There's a reason that not all reviewers are equal. We make quick judgments all the time on how trustworthy a recommendation is based on clues about how it's written. Writing good reviews is hard! That's why there are professionals who do it!

AI-powered reviews will certainly be great for beauty brands that want more high-quality reviews. But they could also be bad news for the beauty buyer who relies on authentic recommendations when choosing which foundation to plunk $40 down on.