The ESRB has added:

“To be perfectly clear: Any images and data used for this process are never stored, used for AI training, used for marketing, or shared with anyone; the only piece of information that is communicated to the company requesting VPC is a “Yes” or “No” determination as to whether the person is over the age of 25.”

Sure, ok…

I don’t know what else to say about this, this will obviously turn into something else.

      • SatanicNotMessianic@lemmy.ml
        link
        fedilink
        arrow-up
        16
        ·
        1 year ago

        From the description, it sounds like you upload a picture, then show a face to a video camera. It’s not like they’re going through FaceID that has anti-spoofing hardware and software. If they’re supporting normal web cams, they can’t check for things like 3d markers

        Based on applications that have rolled out for use cases like police identifying suspects, I would hazard a guess that

        1. It’s not going to work as well as they imply
        2. It’s going to perform comically badly in a multi-ethnic real world scenario with unfortunate headlines following
        3. It will be spoofable.

        I’m betting this will turn out to be a massive waste of resources, but that never stopped something from being adopted. Even the cops had to be banned by several municipalities because they liked being able to identify and catch suspects, even if it was likely to be the wrong person. In one scenario I read about, researchers had to demonstrate that the software the PD was using identified several prominent local politicians as robbery and murder suspects.

            • Shikadi@lemmy.sdf.org
              link
              fedilink
              arrow-up
              3
              ·
              1 year ago

              Yes, it does work like that in some cases. My comment is technically wrong, the best kind of wrong.

              As another commenter pointed out, the way they intend to do it sounds absolutely like they are going to do it the old way, which surprises me because the hold up a photo thing has been a solved problem for a while.

      • Umbrias@beehaw.org
        link
        fedilink
        arrow-up
        3
        ·
        1 year ago

        Plaster sculpt, then add fake skin to, and add a small linear actuator for breathing stimulation, small twitch motors under the skin, and run it under some alternating leds to stimulate blood flow coloration. Should defeat almost all facial recognition software. Might need some eye fakes.

        Or just wear makeup to an insane degree. Or return to the forests and live a much happier life.

      • Radium@sh.itjust.works
        link
        fedilink
        arrow-up
        3
        arrow-down
        4
        ·
        edit-2
        1 year ago

        TI would be every dollar I’ve ever made that you know absolutely nothing about how it works. You seem like someone who is barely technically proficient and likes to pretend like that means they know how things work.

        I’m a software engineer and can confirm that you are absolutely fucking wrong on this one.

        • Shikadi@lemmy.sdf.org
          link
          fedilink
          arrow-up
          4
          arrow-down
          1
          ·
          1 year ago

          I’m a software engineer and I work in machine vision hardware. I may have been lazy with my response, but I do know what I’m talking about. On some level I’m probably in a bubble because I work close enough to the cutting edge of things that I wouldn’t expect any modern company to be employing such basic algorithms to a solved problem.

          I’m a software engineer and I can confirm that you are absolutely fucking rude on this one.