As wonderful as the internet can be, it also lies to us every day. Fake reviews -- seemingly legitimate assessments created by a seller or someone paid by them -- are becoming harder to spot. The online shopping boom has made fake reviews a big business, but they can end up costing you serious money.
In 2021, it was discovered that over 200,000 people were involved in a fake reviews scheme with third-party Amazon vendors. Worse yet, the AI boom has made it more difficult to tell real people from robots. Amazon now has tools that allow sellers to generate product descriptions and create listings and users to ask questions or compare products. The FTC has been empowered to go after fake reviewers, but will that prevent you from falling victim to a scam?
If you can't tell a genuine review from a fraud, we recommend that you consult our comprehensive reviews across multiple categories before making a tech purchase. PCMag reviews over 2,000 tech products a year, and our experts know their markets inside and out. However, there are a few telltale signs a review may not be genuine. And if you really can't tell them apart, some helpful online tools can help clear things up.
What to Look For
As you peruse Amazon, be on the lookout for reviews that provide context for a rating, don't plug competing products, and use natural-sounding language. Reviews that are overly positive or negative without offering enough detail can be suspicious. Very brief five-star and one-star reviews, particularly if they're all posted on the same day, may also denote suspicious activity.
In order to combat fake reviews, Amazon adds a Verified Purchase label to reviews where the reviewer has purchased the product. However, this doesn't take into account schemes where reviewers are compensated for legitimate purchases. If you happen across a review you believe is fake, you can report it to Amazon, and mark it for investigation and possible removal by clicking Report under the review.
Fakespot
Fakespot rates how reliable product pages are on Amazon, Best Buy, eBay, Sephora, and Walmart. Its algorithm looks at reviews and reviewers, analyzing language, previous reviews, and purchase history to determine trustworthiness. Paste the URL of the product page into the Fakespot Analyzer to investigate. For this story, we used the listing for the Amazon Echo (4th Gen). Once the Amazon link is added, Fakespot spits out a grade that should tell you if the page has reliable information.
Fakespot also filters out reviews it believes to be fake to produce a more reliable product rating. Amazon users give the Echo a 4.5-star rating, but after running it through Fakespot, the adjusted rating is closer to 3.5 stars. Fakespot also gives the URL a B review grade, signifying that some of the reviews on this page (which top 147,000) may be somewhat unreliable.
A Fakespot Chrome extension allows online shoppers to analyze web pages with the push of a button, as well as access seller ratings and other benefits. Fakespot also has a browser app for iOS and Android, which can analyze reviews on supported pages right from your phone.
ReviewMeta
Paste an Amazon URL into the ReviewMeta search bar, and the website will eliminate reviews it deems unreliable and replace Amazon's aggregate rating with its own adjusted rating. This tool provides a report card that evaluates certain criteria on a Pass, Fail, or Warn basis. There are also detailed breakdowns -- complete with graphs -- for the factors contributing to the adjusted rating.
Using our test product, ReviewMeta gave the listing a passing grade, but noted that it failed for Suspicious Reviews. This was because 25% of the analyzed reviewers only reviewed this product and 50% of the reviewers had a previous review deleted.
Keep in mind that all on-site assessments are preliminary reports, meaning you can't edit the adjusted rating, not all product reviews are included, and the sampling that is used isn't perfectly random. For a much more complete analysis of the product, enter your email address to get the full report when the analysis is ready.
TheReviewIndex is a simple online search tool focused on aggregating tech product ratings on Amazon, but it can still determine if a product's score has been boosted by fake reviews. Paste the URL into the site and it breaks the product out into different categories based on the words it parsed from reviews. At the same time, TheReviewIndex runs a spam test to ensure that the reviews are authentic, providing a Pass/Fail grade at the end.
Looking at our test product, The Review Index gave it an 8.6 score and a passing spam grade, but only analyzed the most recent 848 reviews. You can then read through pulled clips -- both positive and negative -- that fit under the different product categories, and view the factors that go into spam testing.
Is it AI Generated?
It's possible that AI tools are being used to mass-produce reviews on e-commerce sites like Amazon. They generate text by trying to emulate how humans write, which can lead to some wonky sentence structure and nonsensical ramblings. But how do you tell the difference between AI and a human with terrible writing skills? If a review looks suspicious, run it through one of the many systems that promise to identify AI-generated text.
In our own testing, we used a human-written story and an AI-generated equivalent with GPTZero, Writer AI content detector, and ZeroGPT. These websites let you paste a selection of text, upload a document, or insert a URL. The text is then analyzed, and a score is given based on whether it was written by a human or AI. These tools proved to be accurate, though it's worth noting that results can be less reliable with shorter submissions, such as online reviews.
