Computer software, developed at Cornell, is able to spot fake reviews with almost 90-percent accuracy.
Volunteers were asked to write fake postive reviews to supplement actual certified positive reviews of 20 Chicago hotels -- 800 reviews were tested in total.
After three human scorers weren't able to consistently spot the fake reviews, the software got its turn.
Applying rules, such as truthful reviews typically mention specific aspects of the hotel and decievers use more verbs than nouns, the computer was able to spot the fake reviews with 89.8 percent accuracy.
The researchers are quick to point out that the software only works for positive Chicago hotels for now. In the future, they hope to apply this software to both negative and positive reviews of restaurants and products.