• Ulu-Mulu-no-die@lemm.ee
    link
    fedilink
    English
    arrow-up
    11
    ·
    1 year ago

    Appending reddit to google search has become the only way to get meaningful search results, without it it’s a shitshow of clickbait garbage, I can’t imagine what it will become if it’s not allowed anymore to index reddit data.

    I understand companies not wanting data to be scraped for AI training for free, it’s not only reddit according to the article, also news sites, I think it’s a legit concern.

    I believe at this point governments should wake up and regulate the matter of AI training globally, leaving it to individual companies will only damage users all over the world.

    • athos77@kbin.social
      link
      fedilink
      arrow-up
      9
      ·
      1 year ago

      Interesting thought: Google wants (needs) reddit’s content, and reddit wants to IPO. Why doesn’t Google just buy reddit? It’s pocket change to Google, really, gets them what they want (content), gets reddit what they want (money).

    • Grimy@lemmy.world
      link
      fedilink
      arrow-up
      5
      arrow-down
      2
      ·
      1 year ago

      If you regulate AI, you kill any open source or small time endeavors and turn the whole thing into a shit show. You need vast amounts of data to train models and only a few companies either have it or can afford what they are missing.

      Our whole economy is going to be AI driven soon, google and Microsoft would literally own us.

      I also think Reddit just aggregated that content. Us, the consumer, don’t deserve to get shafted and see AI costs explode just so spez can make a fat pay day off the content we created.

      • Ulu-Mulu-no-die@lemm.ee
        link
        fedilink
        English
        arrow-up
        4
        ·
        1 year ago

        Regulating doesn’t mean blocking, AI needs to be regulated, it should have been already done, look at stuff like deep fakes, some done even with dead people, fakes with actors faces and voices without their consent, and so on, it’s not just about training, it’s also about how the results are effectively used.

        And the fact the training is expensive doesn’t mean everyone should have free reign about it, especially when noone cares about the reliability of the datasets they’re using, of the ethical aspects of it.

        As for reddit, we’ve been already shafted, that’s why we’re on lemmy now.

        • Grimy@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          1 year ago

          You mentioned regulating right after scraping so I thought it pertained to that.

          Also when I say expensive, I mean prohibitively so in a way that creates a soft monopoly. And when you couple that with the very real possibility that AI replaces most desk work in the coming decades, its bleak.

          That being said, I totally agree deepfakes and all that need to be regulated but only on the platforms distributing it imo. Most seem to want to regulate how the technology itself works, gimping it and forcing filters on the user. All of which can really only be done by stopping users from running it locally.

          I think anything other than the lightest touch would be disastrous for both us and the product.

          I’m curious where you would start. I have some thoughts but mainly only a strict opt out policy for individuals.