Should I believe this headline?

  • salarua@sopuli.xyz
    link
    fedilink
    arrow-up
    9
    ·
    edit-2
    6 months ago

    Wikipedian here - AI on Wikipedia is actually nothing new. we’ve had a machine learning model identify malicious edits since 2017, and Cluebot (an ML-powered anti-vandalism bot) has been around for even longer than that.

    even so, this is pretty exciting. from what i gather, this is a transformer model turned on its side; instead of taking textual data and transforming it, it checks to see if two pieces of textual data could reasonably be transformations of each other. used responsibly, this could really help knock out those [dubious] and [failed verification] tags en masse

  • NX2@feddit.de
    link
    fedilink
    arrow-up
    1
    ·
    9 months ago

    Holy fuck my dyslexia tuned “AI tidies” into something really funny

  • Yote.zip@pawb.social
    link
    fedilink
    English
    arrow-up
    1
    ·
    9 months ago

    RE: “should I believe this headline?” I would say yeah this is a reasonable thing to use AI for. I assume they are not going to let it full-auto massacre all Wikipedia citations but as long as they have someone verifying the replacements that the AI is generating then this seems like a semi-auto way to clean up citations. My only worry would be that the AI would become a full replacement for finding sources, in which case people could just start accepting its suggestions as the best answers when manual searching could find a better source.

    The article does say it downranks low-quality sources, but I wonder how often you can type “what I want to be true” into it and have it find a source for nonsense.

  • Chozo@kbin.social
    link
    fedilink
    arrow-up
    1
    ·
    9 months ago

    This is good news. One problem I’ve always had with using Wikipedia as a research source is that while most of the claims may have citations, those citations will often point to dead links, or to pages that may have been updated/edited since the Wikipedia page was originally written and no longer back up the original claims. There’ve been numerous times I’ve seen multiple citations for a single claim on an article, and every single link the citations point to are either dead links or don’t actually say what the claim was, at all.

    Hopefully this helps to clear up a lot of that mess!

    • online@lemmy.ml
      link
      fedilink
      English
      arrow-up
      1
      ·
      9 months ago

      There’s a bot that goes through and identifies link rot so editors have a backlog queue of them to go through.