‘Impossible’ to create AI tools like ChatGPT without copyrighted material, OpenAI says::Pressure grows on artificial intelligence firms over the content used to train their products

  • TwilightVulpine@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    1
    ·
    7 months ago

    Absolutely and I’m with you on that. I think Copyright is excessively long and overly restrictive.

    But that is another conversation.

    The conversation we are having now is how to protect and compensate human creators that need their livelihoods to keep creating in our society as it is, when these new AI tools, trained on their works, are used to deliberately replace them.

    There are many issues with copyright as it is right now, but it is literally the only resort that artists have left in this situation. It’s not a given that opposing copyright hinders corporations. In this particular case there are many corporations salivating at the opportunity to replace human creators with AI, to get faster work, cheaper, to appropriate distinctive styles without needing to hire the people who developed them.

    There is a chilling effect on its own happening here. There are writers and artists today that are seeing their jobs handed to AI, which decide creative works are not a feasible career to have anymore. Not only this is tragic by virtue of human interest alone, since AI relies on human creators to be trained, it’s very possible that they will spiral into recursive derivativeness and become increasingly stale, devoid of fresh ideas and styles.