A Cornell University study demonstrated that humans are really bad at spotting fake reviews. The study participants might as well have tossed a coin. This is why the researchers used machine learning and developed an algorithm that was said to be 90 percent accurate in flagging fake reviews of Chicago area hotels. The researchers called the tool Review Skeptic and made it available to the public.32 The other three prominent apps designed to spot fake reviews are Fakespot, ReviewMeta, and The Review Index. Fakespot examines webpages for their trustworthiness and spits out a reliability grade; ReviewMeta works in similar ways but checks only Amazon reviews. The Review Index likewise determines whether a product score has been inflated by fake reviews.33 Aside from algorithms, the Cornell researchers identified these three tell-tale signs that a review is probably fake to help shoppers base their buying decisions on more solid evidence:

(1) lacks detail: offers general praise instead of specifics; 

(2) includes more first person pronouns: fraudsters resort to a lot of I and me to appear sincere;

(3) contains more verbs than nouns: 

Genuine reviews rely more on nouns than the swindlers who use verbs to write their bogus positive (or negative) stories. In addition, watch out for extremes, either very positive (five-star) or very negative (one star) reviews that don’t provide much detail. 

If several of them were posted at the same time, they are probably unreliable 

Your Task. 

Visit Amazon, Yelp, Tripadvisor, BestBuy, or Walmart. Look up a product or service you might buy or are interested in. Choose a brand that has a number of reviews and critically examine them using the guidelines provided above. To check your accuracy, run your page through one of the apps designed to spot fakes. Compare your results with those of the review robots. Compile your notes and present your findings to the class or in writing if your instructor directs.