The role played by some apps has been picked up by a team led by Professors Daphne Yao and Gang Wan. They have carried out a large-scale study into how apps on Android phones are able to communicate to each other and trade information.
Sometimes app communication is unintentional; sometimes it has been designed in for the purposes of inter-app data sharing, according to Professor Wan. Irrespective of the intention, a significant security breach exists depending on the types of apps a user has installed.
Data mining refers to the computational process of discovering patterns in large data sets. This involves a range of methods, drawing on machine learning, statistics and database systems. Rather than being about extracting actual data, data mining is about extraction of patterns and knowledge from large amounts of data.
The specific threats identified by the researchers were:
Malware apps — specifically designed to launch a cyberattack.
General apps — which allow for collusion and privilege escalation (this may be intentional or unintentional).
The researchers ran various tests on pairs of apps, using a tool they developed called DIALDroid. This was run across multiple devices over a period of 6,340 hours, studying a huge number of apps: 110, 150. Among the apps studied were 100,206 of Google Play¹s most popular apps and 9,994 malware apps from Virus Share. As an example, running a flashlight app, this app was found to work in tandem with a receiver app which divulged the user¹s, geolocation.
According to Professor Yao: “Of the apps we studied, we found thousands of pairs of apps that could potentially leak sensitive phone or personal information and allow unauthorized apps to gain access to privileged data.”
The information about the apps has been presented to the Association for Computing Machinery Asia Computer and Communications Security Conference, which was held in Dubai on April 3, 2017. The researchers are calling for new standards to be put in place to safeguard user security.
