Debates surrounding the implications of artificial intelligence continue to be a topic of interest. In a new development, British researchers have found that by combining machine learning with audio recordings of typing, AI can potentially steal passwords based on the sound of keystrokes. This discovery, shared by Tom’s Guide and BleepingComputer, unveils an “acoustic attack” that can extract data from keystroke sounds with an impressive accuracy of 95%. The researchers conducted an experiment where an iPhone was placed near a MacBook to record the sound of 36 keys being pressed 25 times. Using these recordings, they were able to analyze the acoustic differences between each key by creating waveforms and spectrograms.
To train a deep-learning image classification program called CoAtNet, the researchers utilized these visual representations along with sound, paving the way for information theft. Hackers could potentially record keystrokes using a malware-infected smartphone, or to a slightly lesser extent, via Zoom (93% accuracy) or Skype (91% accuracy). Basic precautions include working in isolated environments, as emphasized in a University of Virginia paper from several decades ago, and managing microphone access across all devices. The researchers suggest that changes in typing style could potentially thwart such attacks, and they recommend additional protective measures such as the use of white noise and randomized passwords with mixed cases. (Read more hacking stories.)
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
