California company develops a machine learning algorithm to detect apps that improperly access users' personal data
Although Android's rules are quite clear, many apps often access personal data that is not necessary for their operation, thus putting users' privacy at risk. Google has decided to improve the fight against these apps, relying on artificial intelligence.
The Californian company has always tried to fight this phenomenon and protect the privacy of its users. There are several apps that, usually after an update, start collecting sensitive information, which has nothing to do with the service they offer. For example, if a navigation app accesses localization, it is normal. The same could not be said if it is a software that is used to read an e-book. Google, in order to find invasive and dangerous apps for privacy, applies a method of comparison, creating groups. Basically, the apps are divided by category and the system analyzes if there is one within the clusters that handles more personal data than the others.
Machine learning to protect privacy
In an effort to make the detection technique even more efficient, Big G has developed an intelligent algorithm that can be used to automatically group various apps and especially to detect those that pose a threat to users' data and their security.
The system makes use of machine learning and helps Mountain View experts unmask dangerous apps. Moreover, as stated in the blogpost with which the company presented the novelty a few days ago, the algorithm will also be used to support programmers to improve their apps as far as privacy and security are concerned.
Google continues, therefore, to make Android safer, not only against malware and viruses, but also trying to identify all those apps that under false pretenses unfairly appropriate users' personal data.