• blindbunny@lemmy.ml
    link
    fedilink
    English
    arrow-up
    166
    arrow-down
    8
    ·
    28 days ago

    This is what happens when a company has no diversity. Most companies dogfood their own production. Reminds me of Google’s gorilla situation…

    • Hubi@lemmy.world
      link
      fedilink
      arrow-up
      70
      arrow-down
      1
      ·
      edit-2
      28 days ago

      This happens even on newer Toyotas, so it’s not exactly company-specific. The issue is the biased training data used for the face recognition system.

      • blindbunny@lemmy.ml
        link
        fedilink
        English
        arrow-up
        23
        arrow-down
        8
        ·
        edit-2
        28 days ago

        This seems more like an excuse. All these companies aren’t using the same training data.

        They literally never tested this on an asian person before selling in the vehicle…

        • Fedizen@lemmy.world
          link
          fedilink
          arrow-up
          73
          ·
          edit-2
          28 days ago

          Toyota is a japanese manufacturer. Likely they localize the feature and the localized version has the problem. Its completely possible they all contract the same software vendors in the US for certification reasons, resulting in similar problems.

        • systemglitch@lemmy.world
          link
          fedilink
          arrow-up
          17
          arrow-down
          5
          ·
          edit-2
          28 days ago

          Your claim is a Japanese company never tested on Asian people? Would you place a bet on those odds?

          • zaph@sh.itjust.works
            link
            fedilink
            arrow-up
            8
            ·
            28 days ago

            Toyota in the US is more American than most American car companies. The tech being different isn’t that big of a stretch.

        • Kaboom@reddthat.com
          link
          fedilink
          arrow-up
          4
          arrow-down
          3
          ·
          28 days ago

          Or potentially it did, found the issue, and werent able to solve before the deadline