Those are the conclusions of researchers from Pennsylvania State University and North Carolina State University, who took a look at the top 1,100 free applications available in the Android Market. They didn’t find anything malicious, but a surprising number of the programs used unique identifiers such as the phone’s IMEI (International Mobile Equipment Identity) number — sometimes without obtaining permission to do so from the user. One concern is that these unique identifiers could be linked to Android users in databases, essentially providing a stealthy way to track what mobile phone users are doing online, similar to the tracking cookies stored by Web browsers. Unlike a tracking cookie, a mobile phone’s IMEI cannot be deleted. The research follows up on work done by some of the same researchers who last year looked at 30 smartphone applications and found widespread sharing of location data and unique identifiers.
Researchers are only now beginning to put together a picture of what’s going on beneath the surface with these mobile phone apps, said William Enck, an assistant professor with North Carolina State University and one of the authors of the study. “I think people are starting to become more aware of this, but I don’t think there is widespread understanding of what the implications are,” he said.
“The paper really expands our understanding of what applications under Android really are doing…. and what they are doing with our data,” said Lee Tien, a staff attorney with the Electronic Frontier Foundation.
The EFF is concerned that these unique identifiers could be used to track consumer’s online activity, but Tien did find some encouraging findings in the study, too. “I was kind of happy to see that there doesn’t seem to be any obvious misuse of the audio video recording capacity for listening in and that sort of thing.”
Enck and his fellow scientists built a program that took the Java bytecode that runs on Android phones and then decompiled it, converting it into something that humans could more easily look at and understand. In total, the researchers analyzed 21 million lines of code. Most of this work was done by computer but the Enck’s team would often go in and manually inspect software that seemed interesting.
“Our analysis uncovered pervasive use/misuse of personal/phone identifiers and deep penetration of advertising and analysis networks,” said the paper, which was presented this week at the Usenix Security Symposium in San Francisco.
The researchers call their work the “initial but not final word on Android application security.”
One of the problems with this kind of analysis that the while it can show what programs are capable of doing, it doesn’t prove that the Android apps are actually using their built in functionality when they are run on mobile phones.
Still, there findings are interesting. More than 22% of the applications the Penn State researchers looked at could send unique identifiers — typically the IMEI identifier — across the network.
Although there are times when programmers might want to actually want to use these unique identifiers — to help police locate a stolen phone, for example — they can easily be misused, and that can lead to serious security problems said Kevin Mahaffey, chief technology officer with mobile phone security software maker Lookout. “Any time you have a unique identifier… people tend to use it for all sorts of crazy purposes, particularly for authentication.”
Verizon is one of those companies using IMEI for authentication, according to M.J. Keith, a security researcher with the Denim Group in San Antonio, Texas. All it takes is an IMEI and phone number in order to access Verizon’s portal for mobile phone users, he said in an interview.
“You can actually use that to reset the portal password,” he said. “You can take over the entire account, change the billing address. You can actually have a phone shipped to you.”
Last year, Lookout reported that one of the application developers cited in the paper — Callmejack — had helped create more than 80 Android wallpaper applications that collected this type of data, sending it to servers in China.
Mahaffey believes that many developers reuse code that’s been written by other developers, and that often this type of data collection may happen without the software maker even realizing that it’s going on. “It puts developers in an interesting place if they’re using opaque third-party code,” he said. “If they have no idea that tracking is going on, it’s very difficult for them to tell users about it.”