If you’ve ever decided to eat at a restaurant because it had good reviews, you’re not alone — 85 percent of consumers trust online reviews as much as a personal recommendation, according to a 2017 consumer review report from BrightLocal. Positive reviews can make 73 percent of customers trust a local business more, according to the report.

But what happens when those reviews were not written by a satisfied customer, but by a machine? Those kind of reviews could be common on sites such as Yelp, TripAdvisor and Amazon, according to research findings from the University of Chicago.

“Misbehaving companies can either try to boost their sales by creating a positive brand image artificially or by generating fake negative reviews about a competitor,” Mika Juuti, a doctoral student at Helsinki’s Aalto University, said in prepared remarks.

Click to resize

Enter the artificial intelligence.

Juuti’s team refined a machine learning model first created by University of Chicago researchers that was capable of generating fake reviews so convincing that “up to 60 percent of the fake reviews were thought to be real” by study participants, according to the university statement.

From there, Juuti “devised a classifier that would be able to spot the fakes,” according to the statement.

“The classifier turned out to perform well, particularly in cases where human evaluators had the most difficulties in telling whether a review is real or not,” according to the statement.

The study was presented at the 2018 European Symposium on Research in Computer Security earlier this month.

Andrew Sheeler: 805-781-7934, @andrewsheeler