• Reddfugee42@lemmy.world
    link
    fedilink
    English
    arrow-up
    18
    arrow-down
    5
    ·
    9 months ago

    Manually attacking your adversary is so 1900s. Train a MASSIVE ai on every classified hacking methodology/procedure document in existence, give it the adversary’s IP ?range, and press GO.

    • jaybone@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      9 months ago

      That would just cause chaos. Which might benefit the attacker but I think you’d be better off with smaller targeted strikes using the secret catalog of CIA/NSA hacking tools already available to them. Or I suppose why not do both.

    • 0xD@infosec.pub
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      9 months ago

      Hacking is too complex to be automated like that - for now at least. Especially the AI buzzword is useless in this context. There are a lot of attempts to automate the process, but none of them come close to manual review by an experienced professional.

      • Reddfugee42@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        9 months ago

        Even if only one in a million succeed, it’s still able to try those million damn fast. And it’s only going to get better logarithmically.