Cross-posted from: https://lemmy.dbzer0.com/post/34117495

[OC]

Original still created by @gedogfx (IG). Title source: “Inkl”

Edit: I’m not on any other social media platforms, so feel free to share this elsewhere if you want

  • Dr. Moose@lemmy.worldEnglish
    7·
    23 hours ago

    It’s incredible how successfully the AI topic has been hijacked by copyright propaganda.

    Copyright is bad, period. Barking at AI for using copyrighted shit for training just makes it less accessible to train for anyone BUT the mega rich.

    Let’s say we do ban copyright data for training. Then only Google, Microsoft and Apple can releastically source data for training. That and countries that don’t respect this like China and Iran right? Ok, so now they hired farms of people to produce training material and release their models that NOBODY can compete with. Everything is literally worse in every possible way now and AI is fully owned by corporate overlords.

    I genuinely don’t understand the though process of these people. We want information to be free and accessible to everyone, no?

    • Starbuncle@lemmy.caEnglish
      5·
      22 hours ago

      IMO it should be illegal to monetize things made by AI trained on copyrighted material unless you also make the model open source and freely available for download. That doesn’t stifle innovation.

      • Dr. Moose@lemmy.worldEnglish
        2·
        22 hours ago

        I think that’s a fair take though the world is allergic to “open source” being a part of our government. I mean, we can’t even open source government projects paid by citizens in most countries so introducing open source as part of our governing process is basically impossible.