Software that spots fake online hotel reviews? Bring it on!

How do you know when to trust hotel reviews posted by users on websites such as TripAdvisor and Yelp? Maybe you think you can spot a fake, but researchers at Cornell University recently developed a software program that they said can sniff out “opinion spam” better than you can.

Researchers led by a grad student used software to test made-up reviews and real reviews of Chicago hotels, according to this Crave blog post. The reviews also were “judged” by real humans.

“The software got it right about 90% of the time,” the blog says. “This is a big improvement over the average person, who can detect fake reviews only about 50% of the time, according to the researchers.”

The example above is straight from the study: Which review do you think is fake?

Laptopmag’s report points to the study’s word choices as giveaways. True reviews used words that apply to the hotel experience (“bathroom,” “check-in” and “price”) while made-up critiques used words connected to a pretext for the visit (“my husband,” “business” and “vacation”).

Researchers say the software needs more testing before it’s ready for prime-time, but it could have a big impact on the crowd-sourcing crowd.

Now back to the example: If you picked the review on the right as the fake, you’re as smart as Cornell’s algorithm.