As Google tries to keep the Play Store ratings free from fake reviews, the company has set up processes to keep content truthful.

Reviews have become an important tool in helping to decide the apps to install. Fake and misleading reviews can erode the trust in the brand, its app and the rating system.

Fake reviews can help to diminish trust in three ways: bad content, fake ratings and ratings that require an incentive.

This year, the Google Play Trust & Safety teams built and deployed processes that combine human intelligence with machine learning to detect and enforce policy violations in ratings and reviews.

At Google, a team of engineers and analysts monitor and study suspicious activities in Google Play’s ratings and reviews to improve the model’s precision and recall, writes Fei Ye, software engineer at Google, and Kazushi Nagayama, “spamologist” at Google, in a post.

The group also asks skilled reviewers to check the decisions made by Google’s models for quality assurance.

In recent weeks, the process identified millions of reviews and ratings that required Google to remove them from the Play Store. The team also identifies and removed thousands of bad apps, due to suspicious reviews and rating activities.

Fake reviews across the web, not only in Google Play, continues to be a major problem. The results from a BrightLocal survey found 33% of consumers believe they spotted lots of fake local business reviews in 2018, up 25% from the prior year. About 89% of consumers 18 to 34, and 74% of those 35 to 54 saw fake reviews this year.

Earlier this month, Amazon deleted thousands of reviews, restricted seller access to customer data on its website, and managed to stop some techniques that trick the site into surfacing products higher in search results, per a Wall Street Journalreport.