Use Your iPhone to Tell What People Are Typing

Researchers at Georgia Tech’s School of Computer Science have seemingly stumbled across a way to turn your iPhone into a spy phone. The technique devised essentially uses the smartphone to detect keystroke patterns. Picking up the sequence of punches, and determining the differing distances between the keys, the iPhone 4 and iPhone 4S can produce an alarming recreation of what is being typed with up to 80% accuracy. Forget about worrying about who just called you – you could be having to worry about who is listening to you while you type.

How have Georgia Tech computer whizzes done this? It begins with the built-in accelerometers that are used to determine the angle the device is being held at in relation to the Earth beneath it. While the iPhone 3GS’s accelerometer was adequate when it came to accurately determining the placement of keys being typed, the accelerometer that came with the iPhone 4 included improvements that made it much more reliable. According to the researchers, most current smartphones include this improvement.

You might be thinking: why not use the phone’s microphone? Well, as it turns out, the microphone samples vibrations in the air at too high of a rate to accurately pick up keystroke noise for one. In addition, most smartphones have built-in security measures that prevent third-parties from gaining unauthorized access to the microphone, whereas accelerometers have no such protection. Therefore, if such technology were to ever be used maliciously, it’d be used using the accelerometer regardless.

But the possibility of hackers actually being able to intentionally determine keystrokes 8 out of 10 times is so far pretty remote. That’s because the formula that researchers used to figure out what was being typed using the iPhone 4 is tricky and certainly not invulnerable to error. The low sample rate and low range of the iPhone accelerometer means that strokes have to be patterned two-at-a-time to deduce a word being typed, and is built atop a series of quasi-translations. Simply put: your average criminal won’t be able to figure out how to do it.

With that said, it’s not your average criminal you need to watch out for when it comes to malware espionage. All it takes is one criminal super genius to suddenly start spying on typists everywhere. There better be good paying jobs waiting for the graduates of Georgia Tech.

Leave a Reply

Your email address will not be published. Required fields are marked *