113
ArsTechnica's response to the AI generated "quotations" (arstechnica.com)
submitted 8 months ago by to c/technology@beehaw.org
58 comments fedilink hide all child comments
 

Backstory here: https://www.404media.co/ars-technica-pulls-article-with-ai-fabricated-quotes-about-ai-generated-article/

Personally I think this is a good response. I hope they stay true to it in the future.

you are viewing a single comment's thread
view the rest of the comments
[–] 10 points 8 months ago (15 children)

He wrote the article himself, he just got mixed up when experimenting with using an AI tool to help him extract quotes from a blog entry. (He is the head AI writer, so learning about these tools is his job.) It was nonetheless his failure to check the quotes he was copying from his note to make sure that he got them right... but an important bit of context is that he had COVID while doing all this. Now, arguably he should have taken sick time off instead of trying to work through it (as he admits), but this would have cost him vacation time, and the fact that he even was forced into making this choice is a systemic problem that is not being sufficiently acknowledged.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 15 child comments
    •
  • [–] 3 points 8 months ago (4 children)

    If only this could have been prevented by, I don't know, not experimenting.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 3 points 8 months ago (3 children)

    It is literally his job to be familiar with this technology, which he cannot do if he does not experiment with it.

    Having said that, doing this experiment while sick was probably ill-advised, so in that sense I agree with you, but in fairness he probably was not thinking clearly while he was sick.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] -4 points 8 months ago (7 children)

    he had COVID while doing all this

    I've had COVID before, it sucks but it doesn't make you stupid.

    he just got mixed up when experimenting

    I don't believe him.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 7 child comments
    •
  • [–] 7 points 8 months ago (5 children)

    Why don't you believe him?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] -2 points 8 months ago (4 children)

    Because it's completely ridiculous. What if we was just phoning it in? He's just going to come out and say it?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 5 points 8 months ago (3 children)

    So in other words, you are just making an assumption.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 5 points 8 months ago (1 child)

    I mean, I find it hard to believe too.

    accidentally used a Chat-GPT paraphrased version of Shambaugh’s blog rather than a direct quote.

    Is quite the umm…. Accident? So the ars writer pulled some junk from ChatGPT, and used it as a direct quote in the article by accident? I suppose maybe depends on how the ars writer takes notes, and maybe should be an opportunity for reflection on how they do that to not have this happen again. Especially since it sounds likely they do ai investigation as part of their main job.

    I mean, the Covid piece is certainly interesting, and sounds like a failure of ars to offer real sick time or a backup writer or something. I’m not sure.

    All around, a big mess. Though, the blog in question was a fascinating read.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 5 points 8 months ago

    I do not see it as being so implausible that he got mixed up due to being sick and thought he was doing one thing while he was actually doing another; as you said, among other things, he should probably reflect on how he takes notes.

    I think that my overall point stands that there is not a good reason not to at least tentatively give him the benefit of the doubt that he screwed up in the way that he said that he screwed up, rather than assuming that he is lying and actually did something significantly worse, especially given that he fully admits to having screwed up and took full responsibility for it.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 4 points 8 months ago

    I don’t believe him.

    I know the internet is full of untrustworthy charlatans, so I can't blame you, but I'm as anti-AI as they come and I do believe him. Mistakes happen, especially in the context of rushed work done while sick. Remember that a lie by a grifter and the truth from an innocent sometimes look exactly the same; effective lies are made off of what was once a truth, after all.

  • permalink
  • fedilink
  • source
  • parent