Where would we be without Google Maps reviews? Knowing the degree of general user satisfaction with a restaurant, an establishment of any kind, an emblematic place and others helps us to be clearer about whether it is worth going there or not. Of course, along the way, we can help companies get more exposure or make it clear that we didn’t like their service.

Google has explained how reviews work on Google Maps, so that we have a better understanding of what happens when they are posted, as there are significant machine learning mechanisms in place to moderate the millions of reviews that end up being received throughout year round.

An AI waiting for what we publish

Google explains that the policy with reviews is updated. They count as an interesting fact that when the COVID-19 crisis arrived, they established in their content policy to eliminate all comments from critics that encouraged non-vaccination and criticism of establishments that followed health criteria. Once this change has been made, the algorithms analyze the review based on these new values.

Like almost everything at Google, machine learning is essential. As soon as you post a review, it is sent to the moderation system. Given the high volume of reviews they receive every day, they use a combined system of machine learning algorithms, along with a human team to moderate.

The AI ​​takes advantage of all the data Google has about you, it wants to know how your account performed before approving your review

The AI ​​is able to analyze its content, examine your account history, to find out if at any time you have engaged in behavior that it considers suspicious. It even assesses your account activity, to know if you made a lot of reviews in a short time, if you moved a lot on social networks or if you tried to attract attention in your Google account to invite someone to vote in a massive way. In other words, they use a lot of the data they have about you.

If the system does not detect a violation of Google’s policy, the notice is published automatically, within seconds. If it is not compliant, the content continues to be analyzed to end up publishing it or finally rejecting it. The human team is responsible for reviewing reported content, whether for SPAM, violence, discrimination and others.

More information | google