(Dmitri “Henry” Aleksandrov, Holder USA) Recent research found this out artificial intelligence could be advanced enough to steal people’s passwords by listening to the sound they type on Zoom and other recording devices.
Academics from the universities of Durham, Surrey and Royal Holloway realized that microphones can detect handwriting patterns, which automatically means that anyone using a laptop in public could have their handwriting recorded and decoded, the daily mail reported.
The researchers formed a AI model pressing all 36 keys on a MacBook Pro 25 times while recording sound. After that, they “fed” the sounds to the AI, which was able to correctly identify the pattern of each key.
In order to prove their theory, the researchers placed a iPhone At 6.7 inches of the same MacBook Pro to record someone typing and the AI was able to deduce the content of the handwriting with 95% accuracy.
The researchers also tested Zoom in the same way, and the AI was able to register keystrokes with 93% accuracy.
“Each key has a unique audio or voice that can be fingerprinted to infer what is being pressed,” said Dr Ehsan Toreini, a researcher at the University of Surrey’s Cyber Security Centre.
A few years ago, academics conducted a similar study on an Enigma machine. The researchers of this study realized that if it had been possible to implant a microphone, a similar result could have been achieved, although the accuracy would be significantly lower: only 70%.
“It gives you a hint of the huge improvement that has happened in the last five years in terms of the accuracy of the models, which somehow raised the accuracy from 70% to about perfect results,” he said Toreini
Then he added this apple should consider adding random noise to keystrokes to prevent “side channel” attacks. In addition, Toreini suggested that technology like Zoom should compress audio to make it harder for AI to steal your private information.