Google claims to be able to remove millions of fake Play Store reviews using AI
A new post on the official blog for Android developers discusses the problem of fake ratings and reviews on the Google Play Store. This activity results in inaccurate and skewed rankings for apps on this Store, according to Google at least. The effects of this activity could also result in fraudulent revenue- or popularity-aggregating on the part of potentially dishonest developers. This, in turn, could undermine the faith users have in the Play Store and the way in which Google approves and recommends apps for download.
Therefore, the company's Play Trust & Safety division has let it be known that they have implemented a new system to 'vet' reviews and ratings for authenticity. The failure to pass these new measures leads to the removal of this feedback from view on the public Play Store. Google claims that their new method for rooting out these 'fakes' incorporates AI trained to tell ratings or reviews that may have been posted by bots from the same posted by genuine human users.
These distinctions are also apparently monitored and validated by "skilled reviewers". The complete system, therefore, filters out potentially fake reviews posted automatically, as well as those that contain other undesirable content such as profanity. It is also designed to address the problem of incentivized reviews, in which users are offered rewards such as paid app versions for free in return for positive reviews or high ratings.
Google appears confident in this new system, to the extent that it now advises developers not to bother resorting to these measures to boost their feedback. The company also now asserts that their new system has already resulted in the removal of millions of ratings and reviews from the Play Store, as well as thousands of apps identified as exploiting the related strategies.