• fiat_lux@kbin.social
    link
    fedilink
    arrow-up
    6
    ·
    edit-2
    1 year ago

    Holy shit, accuracy drops to 93% only when done by Zoom call. That’s very impressive. Privacy truly is dead.

  • Veritas@lemmy.ml
    link
    fedilink
    arrow-up
    7
    arrow-down
    2
    ·
    1 year ago

    The article discusses a new acoustic attack that can steal data from keystrokes with 95% accuracy. The attack works by sampling the audio of keystrokes and using spectrograms to identify the individual keys. While the article mentions that a keylogger is needed to replicate the attack, it raises the possibility that if each key can be identified by sound, it would only require a video of someone writing something to identify other keystrokes using the audio.

    Once this technique is improved to require only a few sentences written on video instead of a keylogger, it could potentially be used to compromise the security of individuals, including streamers.

    • Carbonizer@kbin.social
      link
      fedilink
      arrow-up
      2
      ·
      1 year ago

      There’s plenty of quieter switches, and even some silent ones if you’re that concerned. Gateron/Cherry Reds are quiet enough for most folks, but you could spend a little extra to get a board with Zealios or Zilent switches. Keep in mind that bottoming out the key will still make a lot of noise, but with a bit of conscious practice one can learn to type without it, significantly reducing the noise any mechanical key makes.

  • Obsession@sh.itjust.works
    link
    fedilink
    arrow-up
    2
    ·
    edit-2
    1 year ago

    Given that there’s variance among even the same key switches, as well as differences in sound based on the construction of the keyboard the switches are placed in, the model would probably have to be trained specifically for a given target.

    Since I’m working from home, I have much bigger things to worry about if attackers get access to my physical machine to train their model lol