Today, Google Maps lives largely off its community. Be it the forecast of a traffic jam or the volume of general traffic, the information provided on the points of interest and much more. Without user data and information, Google Maps would be half as attractive and practical. This is why Google keeps asking people to […]
Today, Google Maps lives largely off its community. Be it the forecast of a traffic jam or the volume of general traffic, the information provided on the points of interest and much more. Without user data and information, Google Maps would be half as attractive and practical. This is why Google keeps asking people to participate. Currently, more than 20 million new posts are received every day from Google Maps users. Reviews, ratings, photos and more.
To make our work easier, Google recently integrated a new Contribute tab in the Maps mobile app. Here it is even easier for us to share our information with others.
Google explains the methodology for exposing counterfeits
For example, automated detection systems, including machine learning models, search millions of posts to detect and remove spam and irrelevant content. Our systems check each post before it is posted on Google Maps and look for signs of fake content. Our machine learning models pay attention to certain words or sentences, examine patterns in content that a user account has already posted in the past, and can therefore detect suspicious reviews.
We continually improve our automated systems, but we also know that they are not perfect yet and that fake reviews can happen from time to time. That’s why we also use teams of trained employees and analysts to verify reviews, photos, business profiles, and other content. We also give everyone the opportunity to report reviews, inappropriate content, and misleading places to us so that we can verify their removal.
When we look at the 20 million posts we receive on Google Maps every day (that’s over 7 billion per year), we achieved the following with this method in 2019 alone:
- We have over 75 million reviews that don’t meet our guidelines and we removed 4 million fake business profiles. By improving our machine learning models and automatic detection systems, we can block increasingly trusted content that violates our guidelines. In this way, we also recognize irregularities so that employees can verify them.
- By making it easy for users to report inappropriate content, we have removed more than 580,000 reviews and 258,000 business profiles reported to us, and reviewed and removed more than 10 million photos and 3 million videos that violate our policies. Our teams are getting better and better at intercepting content that violates our guidelines, but is difficult for automated systems to detect (such as inappropriate photos and memes).
- By optimizing our machine learning recognition systems and improving guidelines and training for our employees, we were able to eliminate 475,000 incorrect user accounts.
The new Google Maps for Android is here, with some changes