AI can extract passwords from Zoom typing sounds with 90% accuracy.
A recent study from the UK revealed that by “listening” to an unaware person’s keystrokes, hackers might utilise AI tool to acquire user passwords with nearly perfect accuracy.
An artificial intelligence model was taught by a group of UK-based computer scientists to recognise the noises made by keystrokes on the 2021 MacBook Pro, a “popular off-the-shelf laptop.”
The study results were released by Cornell University, and they showed that when the AI programme was activated on a nearby smartphone, it was able to replicate the inputted password with an astounding 95% accuracy.
The hacker-friendly AI tool also performed incredibly well when “listening” to typing while participating in a Zoom video chat using the laptop’s microphone.
A record for the medium, according to researchers, was set by it in reproducing keystrokes with 95% accuracy.
The researchers cautioned that many users are unaware of the danger that malicious actors could record their typing to breach accounts — a sort of malware they labelled a “acoustic side channel attack.”
The programme was able to “listen” to distinguishing characteristics of each key press, such as sound wavelengths. The iPhone 13 mini was set 17 centimetres away from the keyboard on the desk.
Joshua Harrison from Durham University, Ehsan Toreini from the University of Surrey, and Maryam Mehrnezhad from Royal Holloway University of London carried out the research.
Another risk associated with the developing technology is the potential for AI technologies to assist hackers.
Because AI makes it simpler to tailor schemes to each target, Insider revealed on Tuesday that AI capabilities may make it more difficult to identify online scams.
In 2019, two experts issued a warning about the increased vulnerabilities in internet-connected gadgets that would result from the development of AI and 5G technologies, hence escalating cybersecurity risks.