Meanwhile, readers say that some AI-penned articles switch languages halfway through.

  • AutoTL;DR@lemmings.worldB
    link
    fedilink
    English
    arrow-up
    14
    ·
    1 year ago

    This is the best summary I could come up with:


    Former Gizmodo writer Matías S. Zavia publicly mentioned the layoffs, which took place via video call on August 29, in a social media post.

    Earlier this summer, Gizmodo began publishing AI-generated articles in English without informing or involving its editorial staff.

    The stories were found to contain multiple factual inaccuracies, leading the Gizmodo union to criticize the practice as unethical.

    For Spanish-speaking audiences seeking news about science, technology, and Internet culture, the loss of original reporting from Gizmodo en Español is potentially a major blow.

    Subtle errors, mistranslations, and lack of cultural knowledge can impair the quality of automatically translated content.

    But with so many media companies chasing revenue through SEO manipulations and AI-written filler, it’s unlikely that we’ll see the end of this apparently cost-cutting AI trend soon.


    The original article contains 523 words, the summary contains 129 words. Saved 75%. I’m a bot and I’m open source!

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

            Does it? The summarization library doc link is broken, but it doesn’t mention AI/ML on its page.

            • appel@whiskers.bim.boats
              link
              fedilink
              arrow-up
              1
              ·
              1 year ago

              You’re right, I had a look too and couldn’t find much. The Sumy space is hosted on hugging face though, which does ML stuff. But yea can’t be sure

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

              I had a look at the github repo. The summarization is being done by a library called Sumy.

              After a quick look through it, you’re right that it doesn’t use machine learning. However, it does use a lot of key concepts from Natural Language Processing, such as Tokenization, which is a subfield of Artificial Intelligence.