Google is heavily concerned about vulnerabilities. Since hackings are causing a severe threat to personal info on mobile devices and in fact that is becoming easy to do for hackers, Google got serious enough about Android app privacy.
This time Google’s focus is on trending technology machine learning to tackle with hackings. Yes, Google will leverage machine learning and automated peer review scans to tackle with Android app privacy and secure personal info from app permissions overreach.
Google revealed that it has developed a machine learning algorithm to group apps that are similar in nature. So the algorithm can automatically compare every app’s Android privacy settings and determine if any app requires being deeply inspected by Google’s security and privacy team. Apps are grouped on various metrics like their description, their metadata, the number of installs, and other factors like a set of categories like “productivity” and “games”.
What Google Security Team Says
Martin Pelican of Google’s security and privacy team said that we are concerned about signals that have a negative impact on user privacy, such as permission requests that are beyond core app functionality. For example, a flashlight app doesn’t need access to contact app of the device. Similarly, there are many apps that don’t need access to many functionalities or apps of the device.
In Google’s most recent annual Android security review, it was found that the percentage of users who installed detrimental apps from the Play Store has reduced to 0.05 percent in 2016 from 0.15 percent.
The same review highlights some loopholes in the Android app security system, which happens when users download apps from outside sources. Chinese users download apps from alternative sources, which Google doesn’t have any authority to control. When all this counted together, the percentage of users who downloaded bad apps rose to 0.07%.
In fact, Google is increasingly enthusiastic in implementing AI everywhere in order to build smarter systems, eliminate frauds, enrich customer experience and increase users’ reliability on its products and services. Its acquisition of DeepMind, a British artificial intelligence company, and Google Assistant justify this statement.
Last year (2016), Google used local machine learning to train and improve its Gboard smart keyboard. The technology called Dubbed Federated Learning was leveraged to enhance the user experience with Gboard virtual keyboard while keeping users’ data private. In fact, Gboard gets itself updated with user behavior with it. For example, when the user clicked on its suggestion, the phone locally saves information about the current context, instead of pushing it to a cloud system.
It is really interesting to see that AI-powered technologies are impacting the every area of technology today. Subsequently, IT services companies, particularly mobile app development companies USA and companies that develop mainframe software are drastically shifting their focus to AI to leverage trends. On the other hand, a handful of companies, more particularly forward-thinking businesses are trying AI as a way to achieve a competitive advantage.