Introduction

A recent study warns of the dangers of new technologies that can learn user data by monitoring keyboard typing sounds as they can steal data using microphones in electronic devices.

Risk of Piracy

With the increased use of videoconferencing software like Zoom and the ubiquity of devices with microphones, the risk of piracy has also increased, experts say, according to the British newspaper The Guardian.

Machine Learning Model

A study published during a symposium of an international professional association dedicated to advancing technological progress indicates that researchers have developed a model based on machine learning algorithms that is able to identify the keys being pressed.

The 36 keys on the MacBook Pro were pressed about 25 times with different fingers and with different strokes.

Accuracy of Voice Recordings

The voices were recorded using the Zoom call and on a smartphone located at a short distance from the keyboard, and the accuracy of the voices recorded on the phone reached 95 percent, and the voices recorded using the Zoom reached 93 percent.

While it’s unclear how the keys were identified, Joshua Harrison, lead author of the study at the University of Durham, said it likely depends on how close each key is to the edge of the keyboard.

In turn, Dr. Ihsan Toureni, co-author of the study, said that with the increase in the prevalence of smart devices equipped with microphones inside homes, hacker attacks of this kind confirm the importance of public discussions about artificial intelligence.

The warning came after UK researchers created a deep learning model that can learn the data users enter into their computers by tracking keyboard typing sounds with up to 95% accuracy.