Japan Goes All In: Copyright Doesn't Apply To AI Training
technomancers.ai
Training data can be used "regardless of whether it is for non-profit or commercial purposes, whether it is an act other than reproduction, or whether it is content obtained from illegal sites or otherwise."
      • DarkTides@lemmy.fmhy.mlEnglish
        2·
        1 year ago

        Despite people’s objections the AI revolution is happening anyways too, so its either adapt or be left behind. And pirates will utilize AI tools of their own too if corporate run ones put in stuff like copyright restrictions.

        • Martineski@lemmy.fmhy.mlOPMEnglish
          2·
          1 year ago

          Yeah, copyright will be abolished in the not that far future and it’s awsome. Abolished as in it will make zero sense to have it some point.

  • code_is_speech@lemmy.fmhy.mlEnglish
    2·
    1 year ago

    I’m not a fan of copyright and IP laws in general. But this isn’t the big win it appears at face value. Consumers will still be held liable for copyright and IP violations, while this will absolve the megacorps of any similar responsibility.

    Microsoft can train their AI on 10 million pictures of mario, but when your generated game assets look ‘too similar’ to Nintendo’s, it’s you they will be coming after and not Microsoft.

    There is another major problem, which is in Software:

    Closed source code cannot be used to train AI, but Open Source code can. Open source licenses like GPL have encouraged the proliferation and use of free and open source software and protected against harmful software practices and data collection which can be hidden behind proprietary compiled code. If megacorps are allowed to continue ignoring the GPL license and use it to train AI it could have a massive chilling affect on software freedom and the use of free software licenses.

    Take note that microsoft chooses not to train it’s AI on their own code base, or the code bases of other major tech companies (which Microsoft has access to through its ownership of github).

    Microsoft (and the other major players in AI) know that almost no-one wants their work used to train AI, especially not a mega-corporation’s proprietary, for-profit, AI. So it goes after small timers who can’t afford to fight back.

    Right now, as an average consumer, AI feels great. There aren’t ads or heavy usage restrictions, and the AI isn’t overly manipulative.

    There is no reason to believe that thing’s will stay this way. Right now these tech companies are operating in a massive legal grey area. And they know that the sentiment of the general public will have a massive effect on what they are able to get away with, and what laws get changed. Once these tech companies have pushed the laws through that they need, I have no doubt that AI will change dramatically, with every player looking to monetize as aggressively as possible.

    We need to be very careful about what legal rights we afford these monoliths. And we should be coming up with sensible legal protections now to prevent our manipulation by these tools.