The true cost of generative AI is the erosion of trust.

  • lud@lemm.ee
    link
    fedilink
    English
    arrow-up
    8
    ·
    17 days ago

    Seems a bit hard to verbally exchange keys and sign your speech though.

    • pflanzenregal@lemmy.world
      link
      fedilink
      English
      arrow-up
      4
      ·
      edit-2
      17 days ago

      Yeah, I wouldn’t recommend doing that.

      And that’s not how you establish a PKI or implement AE (authenticated encryption). When I send a voice message e.g. I don’t verbally sign it, that’s actually part of KE (key exchange) protocol.

      Same applies to non encrypted, authenticated communication of course.

      And in case you are talking about physical face to face communication: I think you’d recognize your family and friends without the help of cryptography.

      Edit: if you have any question, feel free to ask! Maybe I misunderstood you? Then I’m sorry

      • lud@lemm.ee
        link
        fedilink
        English
        arrow-up
        2
        ·
        17 days ago

        Of course you don’t verbally sign voice messages. That wouldn’t make any sense at all unless you are a robot or something.

        It was just a joke. But if you wanted to authenticate something (like an email, or voice I guess) in a PKI you would sign it using your certificate private key and the CA would tell the recipient if it’s valid or not.

        • Feathercrown@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          17 days ago

          That’s what the article is about though, voice messages can now be spoofed. It happened to one of my friend’s parents

          • lud@lemm.ee
            link
            fedilink
            English
            arrow-up
            2
            ·
            edit-2
            17 days ago

            I was implying that you do the actual singing verbally. Like actually saying the bits, which would be pretty much impossible.

            Of course you could sign voice messages. That would be like singing any other file.