Researchers in UK said the AI could use video conferencing tools like Zoom to detect which keys are being pressed based on the sound made. In this way hackers could identify laptop passwords and other sensitive information.
In the study the researchers trained machine learning algorithms with sound recordings to create a system that can identify which keys are being pressed with more than 90 percent accuracy, Interesting Engineering reported.
The researchers from the University of Surrey, Durham University, and Royal Holloway, University of London, pressed each of 36 keys on a MacBook Pro 25 times in a row. They varied the fingers and pressure each time.
The typing was recorded both over a Zoom call and on a smartphone near the keyboard.
After training the AI, they found that it could accurately assign the correct key to a sound 95 percent of the time when the recording was made over a phone call and 93 percent of the time when it was made over a Zoom call.
The researchers said their work highlights the need for public awareness and debate on the regulation of AI technology.
They noted ways to reduce the risk of hacking this way, including using biometric passwords or two-step verification systems. The study was published in IEEE European Symposium on Security and Privacy Workshops.