Researchers from British universities have developed an artificial intelligence (AI) model that can steal data from keyboard keystrokes with more than 90% accuracy. This means that the AI program can listen to the sound of a person typing on a keyboard and understand what they are typing, including sensitive information such as passwords and bank account details.
As cyberattacks become increasingly sophisticated with advancements in technology, it is important to stay up to date with the latest hacking methods in order to protect oneself. The AI model developed by the researchers started as an academic study but could potentially become a tool for malicious individuals.
The study, conducted by researchers from Durham University, the University of Surrey, and Royal Holloway University, was published on arXiv. The researchers recorded keystrokes on a respondent’s keyboard to gather training data. They then used spectrograms and waveforms to visualize the differences in the sound produced by each key. By training an image classifier called CoAtNet from Google Brain, the researchers were able to identify keystrokes.
To test the accuracy of the AI model, the researchers placed an iPhone 13 next to the computer to record keystrokes and listened to button presses via Zoom and Skype. The AI model achieved 95% accuracy from smartphone readings, 93% accuracy from Zoom, and 91.7% accuracy from Skype. This technique can be especially effective against people who use mechanical keyboards, which produce audible clicks when buttons are pressed. However, it can also work on membrane-based keyboards.
Fortunately, there are ways to protect oneself from this new AI hacking method. One method is to learn touch typing, which involves typing without looking at the keyboard. By doing so, the accuracy of the AI model can be reduced. Additionally, software that reproduces or filters keystroke sounds can be used, or playing white noise can help muddle the accuracy of the AI program.
It is important to note that this AI hacking method is relatively new and may not be widely used by hackers at this time. However, it is crucial to defend against conventional methods like phishing and scams. Some tips for staying safe online include handling personal information carefully, keeping devices updated, using strong passwords and two-factor authentication, avoiding fake websites and emails, securing internet connections, using password managers, and creating backups of important files.
The development of an AI model that can steal data from keyboard keystrokes highlights the need to stay vigilant against new hacking techniques. By employing the recommended methods for protection and staying informed about the latest hacking methods, individuals can better defend themselves against cyberattacks.