How to Spot a Fake Review

/

Fake online reviews are unfair to competitors and can cause bad purchase decisions. Students benefit from knowing how to spot these reviews, and the topic raises bigger questions about integrity.

People depend on online reviews for purchase decisions, but 30-40% of reviews are considered unreliable. A CNBC article suggests ways to avoid being duped:

  • Beware of five-star reviews, particularly without text or right after a negative review appears.

  • Check reviewers who have locked profiles, post only one time, use a stock photo, or post about businesses in multiple countries.

  • Don’t be fooled by detailed reviews with photos, which can still be fake.

Other sources suggest evaluating the writing:

  • Repeated mentions of the product or brand name

  • Language you would see in an ad or press release

  • Unnatural language (for example, “robust wireless data transmission”)

But some of these strategies require digging and might not be practical for someone perusing Airbnb or looking for a mop. Some research found that language models do a better job than humans in detecting fake reviews, so that might be the best defense.

One tool, Fakespot, a browser add-on, promises to flag fake reviews on Amazon and other sites using AI technology. The Hopeless Geek analyzes tech products and gave Fakespot a mixed review. His point is that the program analyzes reviews, not products. For example, Samsung was flagged as possibly being "deceptive,” but Geek argues the conclusion is based on how the reviews are written. Amazon also protested Fakespot and convinced Apple to remove the app, which it claimed was wrong about products 80% of the time.

Fakespot warns us about the integrity of fake reviews while developing its own credibility as an app. The website uses several strategies that students would recognize to demonstrate ethos; for example, we see “trust” and “truth” several times, reputable company logos (those the app covers and news articles about the app), and testimonials. The site also uses social proof (“millions of consumers like you”) to convince us to install the add-on.

Spotting fake reviews is complicated. If ChatGPT writes a review for someone feeding it the information, is it fake? In a letter, US Public Interest Resource Group (PIRG) encourages the U.S. Federal Trade Commission (FTC) to do more to investigate and monitor websites. Complicated problems require multiple solutions—beyond what individual users can do.