Google Maps Deleted 100M Abusive Edits in 2021 with Machine Learning

While any registered user can add edit on Google Maps, it can be both constructive and destructive. The latter is not meant to improve the usability of the service, and deleting these manually is hard. In steps the AI: machine learning made it possible for Google to roll back about 100 million edits made for destructive or trolling purposes.

 There have been tools for counteracting these changes. For example, users can report any edit, review, or whatever they consider misinformation. There are also moderators that review these reports. But their efforts seemed inefficient to prevent the flood. AI tools seem to be much more powerful.

 What are these abusive edits? Some of them were aimed at businesses that opened after their closure in 2021. Fraudsters impersonated existing businesses in order to take control of their representation on Google Maps by creating profiles before the actual owners or by making false claims. The overall number of these attempts exceeded 12 million. As efficient as these AI-powered tools are, they also require some human assistance to tell, for example, a joke from a real threat. So protein moderators are still required to assist silicon ones.

 In 2022, the problem becomes even more severe. For example, Google blocked any edits on Google Maps in Russia and Ukraine as these edits might be used in war to mark objects for destruction. There were also other ways of (let’s put it this way) using Google Maps not the way it was intended to; for example, users posted pictures and stories of war in Ukraine as reviews to popular places in Russia.

 It took some time to create these reports for 2021. The analytics for 2022, though, will only be available in 2023, and it may report new tools introduced by Google for this. But you can share your views right now in our comments.