The Keys To Acoustic Spying
UC Berkeley released a study on Tuesday highlighting a new security threat to computer users, the sounds of your keyboard. Apparently, your old buddy Qwerty has a language all his own and with the right algorithm, people can determine exactly what you’re typing just by the sounds of the keys.
The researchers took 10 minute recordings of keyboard typing sounds. Then the recordings get plugged into an algorithm and they say they can recover 96% of the characters. They can use this to determine passwords, the story I’m writing now or even the girl on the other end of the webcam you’re trying to put moves on.
"It's a form of acoustical spying that should raise red flags among computer security and privacy experts," said Doug Tygar, UC Berkeley professor of computer science and information management and principal investigator of the study. "If we were able to figure this out, it's likely that people with less honorable intentions can - or have - as well."
Berkeley said they will present their findings on November 10th at the 12th Association for Computing Machinery Conference on Computer and Communications Security in Alexandria, Va.
The key to this process is the sound each keystroke makes when it’s tapped. It makes a distinctive sound. The study says average users go through about 300 characters a minute, leaving time for a computer to isolate the sounds of individual keystrokes and then “categorize the letters based on the statistical characteristics of English text.” The example they mentioned letters “th” happen a bit more often than “tj” and then the word “yet” is going to happen a bit more than “yrg”
"Using statistical learning theory, the computer can categorize the sounds of each key as it's struck and develop a good first guess with an accuracy of 60 percent for characters, and 20 percent for words," said Li Zhuang, a UC Berkeley Ph.D. student in computer science and lead author of the study. "We then use spelling and grammar checks to refine the results, which increased the character accuracy to 70 percent and the word accuracy to 50 percent. The text is somewhat readable at this point."
At this point though, this thing is just getting started. The team says the recording is then played back repeatedly in a feedback loop to train the computer to increase accuracy until no significant improvement is seen. According to the team, three cycles were usually enough to get recovery rates of 88% for words and 96% for characters.
Once the system is trained, recovering the text became more straightforward, even if the text was a password and not an English word. After just 20 attempts, the researchers were able to retrieve 90 percent of five-character passwords, 77 percent of eight-character passwords and 69 percent of 10-character passwords. Scared yet?
They did say there was a caveat to all this high tech spying and that was that they didn’t use Shift, Control, Backspace or Caps Lock keys but did talk about training a program to account for those strokes as well. They said also the mouse hasn’t been figured out yet.
According to Berkeley information, the study builds on a foundation laid by IBM’s Dmitri Asonov and Rakesh Agrawal in which 80% of text was recovered. The biggest difference though was that these two used on typist using the same keyboard for their experiment and that the algorithm they used was trained with known text and corresponding sound samples. Berkeley’s version is much more versatile.
The U.C. Berkeley team varied the conditions a quite a bit more, introducing external factors like music and cell phones going off. Their tool can also “learn” and adapt to various typing patterns.
"Background noise definitely made it harder to recover accurate text, but the differences became smaller after several rounds of feedback," said Tygar. "Given enough tries, the computer algorithm will eventually come up with a pretty good estimate of the text that was typed."
Now you’re still probably not nervous but maybe you should be. Considering the quality of sound recording equipment out here. Bionic ears can pick up noises from quite a distance. Then there’s always the laser microphones. You can lay those on the window and pick things up quite clearly. This definitely reiterates the point for multiple layers of security and while passwords are a start, they certainly don’t solve the problem.
"There are different forms of authentication that could be used, including smart cards, one-time password tokens or biometrics," said Tygar. "That helps with passwords, but it doesn't help protect text documents we would want to keep classified. I'm not sure what the solution is, but it's important that we're aware of this vulnerability."
|
Linear Mode
