• Dyskolos@lemmy.zip
    link
    fedilink
    English
    arrow-up
    57
    arrow-down
    3
    ·
    edit-2
    2 days ago

    I would bet my right nut on the real reason for all this is some AI-billionaire who aggressively pushes this with moneyz. Having every fart we make soon be analyzed by AI is the best “natural” training there could be.

    As a cherry on top is the total surveillance for the state(s). AI will probably do a decent job (despite what the article says) in scanning for potential “threats” to let actual people check.

    But I can’t even comprehend the power that would be needed to actually scan every shit by every person every minute. No data center in the world has this oomph. So it has to be a simple keyword-search (in all possible languages, even leetspeek and co?) To forward to ai. And if ai would just report 0.5% as “suspicious” for manual human control, it would be more supermassive than a black hole. This is just not doable and hence defeats it’s fake reason: protecting the kids.

    So that kinda just leaves ai-training and selective easy surveillances without court-orders. Which also won’t protect kids. As every criminal out there will find a loophole.

    • Tryenjer@lemmy.world
      link
      fedilink
      English
      arrow-up
      4
      ·
      edit-2
      8 hours ago

      Of course, just listen to the CEO of Palantir, he already admitted that that’s his goal. By inference, we can extrapolate that this is the goal of all major business leaders of these companies who are developing AI systems. They need more data to compete with China, and if that requires the West to have authoritarian mass surveillance systems, so be it.

      https://gizmodo.com/palantir-ceo-says-a-surveillance-state-is-preferable-to-china-winning-the-ai-race-2000683144

      • Dyskolos@lemmy.zip
        link
        fedilink
        English
        arrow-up
        2
        ·
        8 hours ago

        While he’s technically not wrong, i hate the world and where it will continue to go to.

    • Chaotic Entropy@feddit.uk
      link
      fedilink
      English
      arrow-up
      11
      ·
      1 day ago

      All the current powers that be, private and governmental, can heartily agree that allowing the public to have any expectation of privacy or autonomy is highly undesirable.

    • REDACTED@infosec.pub
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      3
      ·
      1 day ago

      Can we please stop circlejerking AI into everything? The chat control has been in debate before AI was mainstream

      • Dyskolos@lemmy.zip
        link
        fedilink
        English
        arrow-up
        4
        ·
        13 hours ago

        I had multiple possible reasons. Total surveillance is enough already, the recent aggressive pushing hints towards another added goal.

        You’re free to offer YOUR insight. I don’t even hate AI. I like it.

        • REDACTED@infosec.pub
          link
          fedilink
          English
          arrow-up
          1
          arrow-down
          2
          ·
          9 hours ago

          My insight: EU is not interested in training AI for your corporations, neither are personal chats with likely zero accuracy/factuality good training material, neither is sms-style grammar going to improve any existing AI, everything about this is illogical and pretty stupid. It has always been about control, not… training AI lol

          • Dyskolos@lemmy.zip
            link
            fedilink
            English
            arrow-up
            2
            ·
            9 hours ago

            Illogical? Chat is not just about sms-style dumb texts. It’s images and videos. Trillions of freshly taken photographs. Those are tremendously valuable. And even if it’d be just text, it’s natural training on people. But it’s also video calls, another incredibly valuable thing.

            And sure, the EU has no AI to offer, hence I said “some ai billionaire” or anyone or lobby that wants that shit being pushed hard.

            But as it is just a thought of a possibility I might totally be wrong. As if peasants like us would ever be allowed to know.

            • REDACTED@infosec.pub
              link
              fedilink
              English
              arrow-up
              1
              ·
              edit-2
              8 hours ago

              Again, that is not a good training material. There have been numerous studies on the type of training data we feed and the result of it. This type of content tends to poison the data and lead to equalivent of brainrot for AI’s. This is not very useful data for AI, there are far better sources. Again, seems highly illogical the EU would do all this just to train some shitty AI. Training material should also always be accompanied by context data, which is commonly missing from instant messaging. It’s just too big of a mess.

              • Dyskolos@lemmy.zip
                link
                fedilink
                English
                arrow-up
                1
                ·
                7 hours ago

                Fair points. But “just” surveillance? Anyone worth being surveiled sure wouldn’t be so dumb to use WhatsApp or other stupid crap. I’m worthless to surveillance and even I would not be possible to surveil.

                Just seems weird that it’s pushed so hard. Surveillance was always a must-have, but why now? The moment it gets voted away it’s back on the table.

                • REDACTED@infosec.pub
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  6 hours ago

                  Counter-argument: all my drug dealers use whatsapp. Real life is not movies, criminals are rarely tech savvy.