The operator of the 'X' social media platform has given an undertaking to the High Court to temporarily cease processing the personal data of its European users to train any of its Artificial Intelligence systems.

X are being sued by the Irish data regulator for breaching the GDPR in processing European citizens’ data for Grok AI without an initial opt-out.

The pause is a temporary measure until they return to the Irish High Court.

It finally feels like these data vampires are being tackled even if the pace is a crawl.

  • @[email protected]English
    2•4 months ago

    How can I make sure that with ActivityPub that my comments can’t be legally used to train AI? Does lemmy have some kind of terms of use?

    • DarkThoughts
      12•4 months ago

      They will do it regardless and just pretend they didn’t. GDPR would already protect you if you’re from the EU, but as you can see, they don’t really care about that.

    • just another devAEnglish
      4•4 months ago

      Legally? No idea. What might be adequate protection in the country your instance is hosted, is probably unenforceable in another country where a federated instance might be.

      Technically, you could try by using your own, self hosted instance, and not federating with others, so they won’t be able to scrape your content as easily.

      But realistically speaking, your comments are possibly more likely to be scraped on Lemmy, since it’s so much more open for bots, and your content is replicated to much more servers, not all of which may have noble intents.

      • PikaEnglish
        1•4 months ago

        That’s true as well that I hadn’t thought of, one could just spin up a VPS in a country that doesn’t give a shit about licensing agreements and use a bot to scrape the data, or even just spin a instance up and Federate/subscribe with everyone that is an open instance / doesn’t use the whitelist (i.e most of the instances) and scrape that way. The only real defense against the latter is defederation, but that system isn’t very advanced and is easily bypassable if you don’t have a bunch of users

        Personally I would be absolutely shocked if this wasn’t already happening, downside of a free/decentralized protocol I guess

        • just another devAEnglish
          1•4 months ago

          Better yet, just spin up your own instance, subscribe to all major communities, and have the servers push the comments to yours. No scraping required, and nobody will ever find out it was you.

          Statistically it’s likely to have happened already.

    • PikaEnglish
      2•4 months ago

      I don’t think you can, some have a license notice at the end of each comment but I have major doubts that it would actually hold up in any actual court hearing