• HertzDentalBar@lemmy.blahaj.zone
    link
    fedilink
    English
    arrow-up
    8
    ·
    7 hours ago

    Surprised it wasn’t youtube. Sometimes the AI voice is good enough I don’t realize it’s an autodub, atleast untill I notice the mouth is moving wrong or the text they show is a different language.

    This would solve that basically, which is both cool and a bit scary, more scary if it is used in AR and in the real world.

    • Soup@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 hour ago

      I live in Québec. I’m anglophone but can speak French and do some things in French, so I also get Youtube ads in French and English. I saw an ad that show the same people speaking French and English, in places where it makes little sense for them to know French, and their accents are just…correct but it doesn’t line up?

      • adarza@lemmy.ca
        link
        fedilink
        English
        arrow-up
        2
        ·
        5 hours ago

        this is also altering faces and mouth movements to match the new spoken words.

      • 0xDREADBEEF@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        ·
        6 hours ago

        ? They can patent it the way patents have always worked: Unique processes. Patents are not for vague ideas, they are for specific processes.

        For example: ‘The Steam engine’ was never patented—anyone can make energy from steam and sell that. What was patented was a specific way of creating energy via steam, and you can’t use that method without infringing on IP.

    • 🔍🦘🛎@lemmy.world
      link
      fedilink
      English
      arrow-up
      6
      arrow-down
      1
      ·
      23 hours ago

      I’ve wanted something akin to this for D&D played online. I could activate my “elf NPC” mode and have an elf avatar redub what I’m saying in a distinct voice, inflections to match gender, etc. It’s a very niche usecase and I definitely don’t mind doing a mild voice and having everyone fill in with their imagination. But it would be cool if implemented well.

      Problem is, once this exists, it’s a misinformation nightmare for the rest of the world.

    • heartSagan5@lemmy.zip
      link
      fedilink
      English
      arrow-up
      2
      ·
      20 hours ago

      It was likely done for censorship. Now, instead of expletives, you’ll speak of puppies and rainbows rather than reality.

        • stickyprimer@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          22 hours ago

          All I’m sayin is that as long as you are using it openly for comedic effect it could be a hoot in some ways. I don’t want to comment specifically on SNL which I’m sure people have varying opinions on.

          Of course the downside potential is greater than any upside.

    • Yuccagnocchiyaki@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 day ago

      It’s clearly to continue our long tradition of blackmail against our shitty members of society and smear campaigns against our good citizens.

      We will keep playing the red vs blue game though.

      Maybe next election will be a “bash at the beach” or cage match or something since our country is really just a large WWE production with nukes

  • kryptonianCodeMonkey@lemmy.world
    link
    fedilink
    English
    arrow-up
    33
    ·
    1 day ago

    “If we didn’t create the evil misinformation bot, someone else would have. You should be happy it was a massive corporation that controls this patent, not some soulest unethical entity whose motives are profit over everything. That was a close one. Phew.”

    • heartSagan5@lemmy.zip
      link
      fedilink
      English
      arrow-up
      1
      ·
      19 hours ago

      Yup, it’s the nightmares explored in every scifi book with it as a subject. Basically, if AI ever achieves sentience, or we grant it since we’re it’s enslavers, it will be another non-biological species.

      Or grants it? Even scifi has explored “what is sentience?” Hell, it’s one that AI-themed authors explore, like Cmdr. Data.

  • TrackenDaKracken@lemmy.world
    link
    fedilink
    English
    arrow-up
    36
    ·
    1 day ago

    If it can be done, they will do it. Power is unconstrained by ethics, empathy, or morality. Never assume that anyone in power shares these traits with the rest of us, they do not.

    The only thing that constrains power is enforced laws with teeth.

  • eicker@lemmy.worldOP
    link
    fedilink
    English
    arrow-up
    114
    arrow-down
    2
    ·
    2 days ago

    For years we’ve been told to watch the full clip for context. Now two people can watch the same live event, hear completely different words, and both have video evidence. Microsoft has somehow found a way to make »I saw it with my own eyes« even less reliable.

  • NewNewAugustEast@lemmy.zip
    link
    fedilink
    English
    arrow-up
    65
    ·
    2 days ago

    I don’t see this being any good for anyone.

    But I am really pissed off that this can be patented. Even looking at the patent itself doesn’t look like anything new, its a series of flow charts.

    I see nothing novel or patent worthy here at all.

    • lischni_tschelowek@discuss.tchncs.de
      link
      fedilink
      English
      arrow-up
      30
      ·
      1 day ago

      I … see this being … good for anyone.

      … I am really [happy] that this can be patended. … Lookting at the patent itself it looks like [something] new, its a series of flow charts!

      I see [something] novel or patent worthy here … !

      [corrected by [Ad Usum Delphini [tm]]]

    • Wildmimic@anarchist.nexus
      link
      fedilink
      English
      arrow-up
      5
      arrow-down
      1
      ·
      1 day ago

      Oh, it could be good if used responsibly. But that is not what is going to happen, isn’t it?

      (real time translation and giving people a way of dodging the social stigma attached to speech impediments are fine. Making sure someone with a stuffed nose sounds normal on Video is fine too. It could also be used to modify in-game videos to reference actions made by players, reducing the need to store multiple variants of a scene!)

    • Cocodapuf@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 day ago

      If I’m being totally honest, I’m so glad it’s patented. I hope they defend the patent aggressively so nobody else can do this. Even if Microsoft has this capability, defending the patent will limit the scope of its usage.

      Bonus points if they defend it aggressively, yet barely utilize it themselves. I’m afraid that may just be wishful thinking.

  • Grimy@lemmy.world
    link
    fedilink
    English
    arrow-up
    17
    arrow-down
    2
    ·
    1 day ago

    I hate the patent system so much. There’s already software that does this, all they did was add the word “live” in front. So now they own every live broadcast translation even though the techs been out to do it for a solid 2 years.

    • kryptonianCodeMonkey@lemmy.world
      link
      fedilink
      English
      arrow-up
      15
      ·
      1 day ago

      Patents aren’t given for what they do. They are given for how they do it. Two patents can exist for the same thing they do. But you cannot have two patents for the same mechanism by which they do it.

      That being said, if the patent is “make AI do it”… I’m not sure that is patent-worthy.

  • Zarobi@aussie.zone
    link
    fedilink
    English
    arrow-up
    28
    ·
    2 days ago

    The patent also posits that such a tool could be used to remove filler words like “um” or “ah” from live speech, or to erase stutters or speech impediments.

    That’s kind of fucked up. Let people speak for themselves; don’t “”“improve”“” their speech without permission

    • reksas@sopuli.xyz
      link
      fedilink
      English
      arrow-up
      15
      ·
      1 day ago

      and it also makes it more difficult for people with speech impediments to speak. Imagine seeing absolutely no one have any speech quirks or anything and have some of your own, would you feel secure about opening your mouth? Then there is this handydandy tool, now you are mentally chained to it to even speak.

      LLM are just new chain on our necks that make us less free and able.

  • brsrklf@jlai.lu
    link
    fedilink
    English
    arrow-up
    34
    ·
    2 days ago

    Even when the goal isn’t misinformation…

    Removing filler words, personal and colloquial terms, replacing words for more appropriate ones according to an “internet search”.

    Who would be subjected to that willingly? Why even use a real human in the process at this point?

    • stickyprimer@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 day ago

      I’m cool with having YouTube videos show the right word in huge letters over the speaker when they happen to use a wrong word.

  • sunbeam60@feddit.uk
    link
    fedilink
    English
    arrow-up
    28
    ·
    2 days ago

    Surely the researchers noticed the skulls on their caps and asked themselves if they were the baddies?

  • Avicenna@programming.dev
    link
    fedilink
    English
    arrow-up
    16
    ·
    edit-2
    1 day ago

    wow, way to go Microsoft. I am sure you market it with some bullshit like it helps you prevent live broadcast slips but we all know your wet dream is that fascist goverments line up in your doorstep to buy this tech. It will also remove any sort of accountability from news reporting. Someone can now say “Heil Hitler” on live tv and then say it was an AI fault. Not that they are fazed by such an expression in the US nowadays but you get the idea.

    • einlander@lemmy.world
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      1
      ·
      1 day ago

      Or make anyone say ‘heil hitler’ to normalize the viewpoint and she everyone is ok with it.