JAKARTA - Google Maps, one of the most often needed applications to navigate roads is often filled with false information, and on average many scammers are trying to bring down businesses.

The app is not only about navigation or exploring the world, there are also contributing content, such as videos, reviews, and ratings that can help users find the best restaurants in the city or other businesses.

Contributed content is indeed very helpful, but sometimes scammers post scam content for money profits.

The company is not standing still seeing this, with the help of a machine learning model to uncover new patterns of abuse, which can work automatically and manually advanced, Google has successfully blocked or removed more than 115 million reviews that violate policy, with most of these reviews known before being viewed in 2022.

Thanks to the updated model, Google has also removed more than 20 percent of the fake reviews than we did in 2021.

"Our techniques block or remove more than 200 million photos and 7 million blurry, low-quality, or violating our content policy," said Ashish Gupta, director of Engineering, User-made Content, quoted from Google's official blog, Monday, April 3.

Gupta added that the automated system detected a sudden increase in the Business Profile with a website that ended with a.design or.top something that would be difficult to spot manually in millions of profiles.

"Our team of analysts quickly confirmed that this website is fake and we can delete it and deactivate the related account quickly," Gupta said.

In addition, Google has also suspended 20 million attempts to create a fake Business Profile, which is an increase of 8 million compared to 2021. Protection for more than 185,000 businesses after detecting suspicious activity and abuse efforts has also been implemented by the company.

Furthermore, the tech giant also shared a way that scammers do to cheat on Google Maps. They overlay inaccurate phone numbers over donated photos, in the hope of tricking suspicious victims into calling scammers instead of actual businesses.

Google stated that its new machine learning model can recognize the numbers left on the contribution images by analyzing specific visual details and photo layouts.

"With this model, we managed to detect and block most of the scams and policy violations before they were published," Gupta concluded.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)