Security researchers from five US universities have teamed up to demonstrate how motion sensors can be used to spy on Android phones.called EarSpy is a proof-of-concept security scare piggybacking on the vibrations produced by your phone’s ear speaker, detected by a motion sensor.
The measurements allowed scientists to learn personal characteristics such as voice, caller identification and gender simply by reading the motion sensor data. This has been done before, but in phone loudspeakers, with the advent of stereo sound in modern phones, earbuds are getting more and more powerful, and researchers are applying machine learning algorithms to I was able to eavesdrop on the owner and collect motion sensor data.
To prove that phone makers need to equalize sound pressure during phone conversations, and that motion sensor placement should make earbud vibrations imperceptible, they’re old I used 2016. OnePlus 3T that didn’t really show up in the spectrogram, stereo sound in 2019 OnePlus 7T registered and later The OnePlus 9 series also did poorly.
OnePlus 9 EarSpy Distribution
However, this so-called side-channel attack can be easily thwarted simply by turning down the volume on your earphones. This makes it harder for motion sensors to pick up viable reverberations that may later be attributed to gender, caller identification, or private speech.
With Android 13, Google is trying to prevent such privacy backdoors by asking permission to collect sensor data at a sampling rate of 200 Hz or higher. However, at 200 Hz, the accuracy of the selected signal drops by only 10%, so it’s not really a preventative measure, but at the standard 400-500 Hz sampling rate, the motion sensor response to speech recognition is nearly impossible.
“Since there are 10 different classes here, the accuracy is 5 times better than random guessing. This means that the vibrations caused by the ear speaker caused a significant amount of discernible effect on the accelerometer data.‘ warns researchers.