The vulnerabilities that artificial intelligence can create in the field of cyber security are scaring researchers. A recent study conducted in the US showed that artificial intelligence can guess passwords almost 100 percent accurately by listening to keystrokes on the keyboard.
How to crack passwords using AI?
A new study from Cornell University looks at the predictions AI can make with keystrokes. It was stated that a password attack carried out by hackers using artificial intelligence could make 95 percent accurate predictions.
Researchers trained an AI model called CoAtNet to identify keystrokes based on sounds from a smartphone microphone. The model achieved 95 percent accuracy in reproducing keystrokes typed on a MacBook Pro keyboard.
Even in noisy situations, such as Zoom calls and Skype conversations, the model was over 90 percent accurate, the researchers said. The volume of the keystrokes also had almost no effect on the accuracy of the password guesses.
According to the researchers, the AI used the timing between keystrokes and the wave of sound coming into the microphone to identify keystrokes on the keyboard. This means that even silent keyboards are vulnerable to attack.
The results of the research are worrying for personal data privacy. Applications developed by hackers for voice eavesdropping could compromise millions of users’ passwords with the help of AI.
Experts recommend using password managers and biometric authentication methods such as fingerprints to reduce the risk.
So what do you think about this issue? You can share your opinions with us in the comments section.