(916) 972-9000 info@connectech.us

It may sound like something straight out of a science fiction movie, but recently, researchers have made a disturbing discovery. Using nothing more than an off-the-shelf microphone, it’s possible for an attacker to determine what content you’re viewing on your computer monitor.

The researchers tested a variety of LCD screens (with both LED and CCFL backlighting) and observed that the high-pitched noise made by the monitors changed as the content displayed on the screen changed.

The research team tested a variety of methods of recording audio data from the monitors in question, and found that they could capture sufficient data with a few methods. First, a smartphone positioned near the screen, second a compromised smart virtual assistant (like Google Home or Amazon’s Alexa), and third, using a parabolic microphone from up to ten meters distant.

Even more disturbing, the researchers discovered that an attacker could correctly identify the website a victim was looking at with up to 97 percent accuracy if the microphone was close to the monitor, and with 90.9 percent accuracy with microphones placed at some distance.

Worst of all, subtle changes in the pitch of your display screen make it possible for hackers to identify what specific characters are being displayed with an accuracy that ranged from 88 percent (more distant microphones) to 98 percent (microphones in close proximity to the monitor). This makes it entirely possible to glean passwords and other sensitive information.

Granted, this is an extremely exotic form of attack that requires a great deal of advance preparation by the attacker. Because of this, it’s unlikely that it will see widespread use anytime soon. Even so, it’s something that a careful and determined hacker could do right now using off the shelf technology, and there’s very little the victim could do to prevent it.

While we’re unlikely to see equipment manufacturers take the necessary steps to mask acoustic emanations, robust on-site physical security measures would make detection of this type of attack fairly easy.