Google introduced Google Play Protect at its I/O 2017 developer conference. The platform brought in a comprehensive set of security standards that helps Android find malicious apps and take actions against them. Google accounts a lot of that to AI and machine learning. For that, in a recent blogpost, Google talks about how the addition of the technology has helped strengthen the security of the platform, and made it possible to keep 2 billion Android users safe.
One of the key operations of Google Play Protect is to look for PHAs (potentially harmful apps) in the Android ecosystem. For this, apps are rigorously analyzed before they show up on Google Play Store. This process itself allows Google Play store to be nine times more secure than before. It doesn’t just stop there. Even after being installed, Google Play Protect constantly keeps a check on them for suspicious behavior. What’s more impressive is that Google’s security system manages to scan 50 billion apps a day to search for malware.
Machine learning comes in to play here. Using its huge applications catalogue, Google has developed algorithms that can help the system identify potentially dangerous apps from the list of safe apps. Based on the understanding of these algorithms, the security system will able to compare and find new PHAs that show similar behavior.
When PHAs are identified, they are put in a category of similar applications. If a threat is indeed detected, necessary actions are taken by Google Play Protect and the information is fed back to the system in the form algorithms, so that it can learn and identify similar PHAs in the system.
Watch: Moto Z2 Force Review
With machine learning and AI, Google has been able to cutdown malware threats by 60.3 percent in 2017. Google plans on increasing that number this year by putting more computing power and improving on machine learning algorithms.