A recently developed AI tool has raised concerns about the security of passwords entered during activities like Zoom calls. This new tool can extract passwords from the sounds of keystrokes, potentially putting users at risk of cyberattacks. The AI has demonstrated an impressive accuracy rate of about 90% in its ability to decode passwords from these auditory cues.
Recent research has unveiled a startling vulnerability that arises when users type in their passwords while participating in online activities such as Zoom calls. Hackers can now exploit an artificial intelligence (AI) tool that “listens” to the sounds of keystrokes, giving them the ability to steal passwords.
The study, conducted by researchers from Durham, Surrey, and Royal Holloway universities, has highlighted that AI tools can differentiate between specific keystrokes by analyzing the sounds they produce. The research suggests that microphones can detect unique typing patterns that are associated with individual keys.
The research, made public on August 3rd, underscores the rising threat of sound-based cyberattacks due to the increasing popularity of video conferencing platforms like Zoom. With the widespread use of devices equipped with microphones, the opportunity for such attacks has multiplied.
The AI tool studied by these researchers has demonstrated an exceptional aptitude for deciphering text, including passwords, from the audio of keystrokes. The accuracy of this tool surpasses 90%, indicating its potential effectiveness for unauthorized password extraction.
The research team accomplished this feat by training an AI model through a meticulous process. They pressed each of the 36 keys on a MacBook Pro 25 times while recording the corresponding sounds. This sound data was then fed into the AI model, enabling it to recognize the distinctive auditory pattern associated with each key.
The AI model devised by the researchers exhibited an impressive accuracy of 93% in interpreting keystrokes from audio recordings of typing on a MacBook via Zoom. Remarkably, the accuracy rate increased even further to 95% when the keystrokes were captured using an iPhone 13 mini.
This research serves as a wake-up call for individuals and organizations that heavily rely on online communication and collaboration tools. While the convenience of remote interactions is undeniable, it is crucial to recognize the potential security risks associated with them. Practicing caution while entering sensitive information, such as passwords, during online activities is now more important than ever.
Read more: Wi-Fi Balloons: Google’s Plan to Provide Global Internet Access
The advent of AI tools capable of extracting passwords from keystroke sounds emphasizes the need for heightened cybersecurity awareness. Users must remain vigilant and consider adopting additional security measures to mitigate the risks posed by such emerging technologies.
