It may be impossible to defend cell phones against being used as bugs because it is difficult to tell legitimate applications that need to use the phones' cameras and microphones from those that abuse them, researchers say.
5 lab technologies that could reinvent cell phones
According to a paper presented at the recent Sigcomm conference in Barcelona, the challenge is great but the researchers from the University of California at Davis and Sprint say there is hope.
"[W]e believe that it is very challenging, if not infeasible, to eliminate the sensor-sniffing threat completely," the researchers said. "However, we are cautiously optimistic that a user-friendly solution, which works effectively in most circumstances, can be devised."
Information can leak out via many technologies , but cell phones are unique because they have sensors that are supposed to pick up activity and conversation. The problem is for the phones to tell spy applications from legitimate ones based on their behavior.
Some applications, like making a phone call, turn the microphone on and off as needed for its main function. Others, such as voice dialing, use microphones as a tool toward placing calls.
In those cases, users are aware the microphones are turned on. But other applications may use the microphone or camera without the user being aware. For instance, an application may turn a microphone on to measure how noisy it is around the phone and adjust the ringtone so it can be heard.
Spy applications could act just like any of these three classes of legitimate phone applications. Applications could be certified to be safe, but the standards set for them might differ from what users would want. And users might be tricked into downloading malicious software anyway, particularly if it is bundled with compelling software such as a game, the researchers say.
Requiring applications to list all the privileges they need to function could help users weed out malicious software, but it would require diligence by the user -- something the researchers say can't be counted on.
An effective defense must prevent malicious programs from accessing phone sensors, require no user intervention, require no modification of existing applications, should not degrade performance and should be applicable to the wide range of mobile devices.
Researchers have considered white-listing and black-listing applications, but note that users likely don't have the expertise to determine if an application should be trusted. They have also considered notifying users with a flashing icon whenever a microphone or camera are in use, but say that such warnings are easily overlooked.
Phones could use their own features to enhance defenses. For instance, use of phone sensors could be prohibited in certain physical locations as detected by the phone's GPS. Also, if a microphone is turned on, for instance, but no spoken word follows in a short interval, the phone could conclude that the microphone is being abused and shut it off. Both these methods are imprecise, the researchers said, and could block legitimate activity.