Google Play Protect is Google’s answer to the ever-evolving battle against malware. We saw it launch in May of last year, and since then it has undergone many changes and improvements. It also rolled out to many more devices following its initial testing phase. Play Protect works by scanning applications uploaded to the Google Play Store for any features designed to harm your device. It also scans applications from non-Google Play sources. It’s constantly improving, and Google’s latest “Android Security Year in Review” aims to prove that.
According to Google, 60.3% of Potentially Harmful Apps (PHAs) were detected via machine learning algorithms, an impressive number which Google only expects to rise. Google is no stranger to machine learning. The company has been integrating machine learning algorithms in every place they can for a couple of years now. Machine learning allows Play Protect to scan as many as 50 billion applications a day while also scanning devices themselves, even when they are offline. 39 million PHAs were prevented from being installed thanks to Google Play Protect. Offline scanning has contributed to another 10 million PHA detection, as 35% of such applications are installed when a device is offline.
As a result of these measures undertaken by Google, those who installed applications only from the Google Play Store were 9 times less likely to get a PHA than those who installed from other sources. The service reviewed 65% more applications in 2017 when compared to 2016. Thanks to the fact it also protects Android’s ecosystem as a whole, there was a 60% drop in installation rates of PHAs from outside of the Google Play Store. This is a huge drop and shows just how well Google Play Protect is currently working.
Of course, we’ve seen some false positives along the way (though this is partially the fault of the OEM), but overall the tool is beneficial. With numerous exploits existing in the wild, Android security patches and Google Play Protect are tantamount to security.
Source: Google Security Blog
from xda-developers http://ift.tt/2FLBRqM
via IFTTT
Aucun commentaire:
Enregistrer un commentaire