Over 170 million fake reviews removed from Google Maps by new algorithm

It turns out that a new machine learning algorithm introduced by Google has removed a large number of fake reviews that were rampant on Google Maps.

How machine learning keeps contributing content helpful


Over 170 million fake reviews were removed from Maps and Search thanks to Google's new algorithm | TechSpot


According to Google, reviews posted on Google Maps are screened by a moderation system, and even after the review is completed and published, they are monitored over a long period of time to see if there is any suspicious activity.

Google has introduced a new machine learning algorithm that detects suspicious review patterns faster than before. This algorithm monitors reviews and the accounts of reviewers, and detects whether a reviewer has left the same review for multiple services or whether there has been a sudden spike in 1-star or 5-star reviews for a certain service. It detects signals and provides information to help identify suspicious reviews.

By introducing this algorithm, Google seems to have succeeded in deleting more than 170 million fake reviews in 2023, 45% more than the previous year.

An example of using machine learning algorithms is to analyze reports from merchants that have seen a recent surge in fake 5-star reviews, identify patterns such as what categories of businesses are being targeted, and identify malicious fake reviews. Among other things, he helped protect businesses that were targeted by review intermediaries.

In this case, Google's research team uncovered more than 5 million fake review attempts in just a few weeks, disabled the associated malicious accounts, and applied protections to hundreds of businesses.

Dan Pritchett, Principal Software Engineer at Google Maps, said: 'To maintain the accuracy of information posted to Google Maps in a rapidly changing world, we use a combination of technology and skilled operators to protect against bad actors. We regularly work with a wide range of industries to fight the fight.”

in Web Service, Posted by log1p_kr