Is Meta Scraping the Fediverse for AI? - We Distribute
wedistribute.org
Is a large corporate entity scraping a community-run open social network to train AI models for profit?

Building on some initial reports coming from the FediPact account and Dropsite news, we dive into potential measures admins can take for their instances.

  • ramble81@lemmy.zipEnglish
    281·
    2 days ago

    Every time this pops up I have the same thing to say… there is nothing that is stopping them from setting up their own federated instance and via the ActivityPub protocol have everything delivered to them in a neatly formatted package ready to ingest, no scraping needed and nothing we could do except try to defederate with them, but we’d have to know which servers are theirs.

    • Zaktor@sopuli.xyzEnglish
      2·
      2 days ago

      I’m more upset that they’d be scraping the HTML rather than just federating and saving the server bandwidth.

      • ramble81@lemmy.zipEnglish
        2·
        2 days ago

        Yeah I understand the resource utilization concern but a lot of people are pissed about ingesting their comments. There were people who actually thought putting CC terms on their posts would actually do anything.