Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
AI The Media

Gizmodo and Kotaku Staff Furious After Owner Announces Test of AI Content (futurism.com) 159

Futurism reports: G/O Media, a major online media company that runs publications including Gizmodo, Kotaku, Quartz, Jezebel, [the Onion], and Deadspin, has announced that it will begin a "modest test" of AI content on its sites... In an email to staff, G/O Media editorial director Merrill Brown argued that the news shouldn't come as a surprise since "everyone in the media business" has been considering AI.

The trial will include "producing just a handful of stories for most of our sites that are basically built around lists and data," Brown wrote. "These features aren't replacing work currently being done by writers and editors, and we hope that over time if we get these forms of content right and produced at scale, AI will, via search and promotion, help us grow our audience..."

Unions representing G/O Media and The Onion staff issued a statement, writing that "we are appalled by this news. The hard work of journalists cannot be replaced by unreliable AI programs notorious for creating falsehoods and plagiarizing the work of real writers." Gizmodo and Kotaku staff, in particular, were outraged at the news. "AI content will not replace my work — but it will devalue it, place undue burden on editors, destroy the credibility of my outlet, and further frustrate our audience," Gizmodo journalist Lin Codega tweeted in response to the news. "AI in any form, only undermines our mission, demoralizes our reporters, and degrades our audience's trust."

This discussion has been archived. No new comments can be posted.

Gizmodo and Kotaku Staff Furious After Owner Announces Test of AI Content

Comments Filter:
  • Good (Score:5, Insightful)

    by NomDeAlias ( 10449224 ) on Saturday July 01, 2023 @08:48PM (#63649868)
    The AI written articles will probably be better.
    • Re:Good (Score:5, Interesting)

      by ArchieBunker ( 132337 ) on Saturday July 01, 2023 @09:25PM (#63649936)

      I read jalopnik on occasion and the batch of writers they have seem to be new to this whole car thing. One wrote a series of articles about a pedal bike they found and wanted to restore. The other bought a basket case Mercedes 190E without ever having touched a wrench. Half the content is the site asking a question and then posting people’s answers as a slideshow in a few days. They used to have knowledgeable writers with interesting content but that was over a decade ago.

    • I think it is fair to say that the desire to elevate these fringe review blogs into something intellectually stimulating failed and was abandoned a decade ago at least.

      What we have left is reprocessed drivel in most cases... if Skynet wants to churn that out... might as well.

    • The pen was mightier than the sword, then the liberal arts majors came for the STEM nerds and they got automated.
      • Re: (Score:3, Informative)

        by Teresita ( 982888 )
        Yeah, one time I asked ChatGPT to give me the distance to the second Earth-Moon Lagrange libration point (L2), it gave me the figure for the Earth-Sun system. I politely pointed out the error and the damn thing acted like a Trump voter being told the orange goober lost the election. So STEM nerds are safe for now.
    • There's a lot of 'content' on the Internet these days that's more or less the same recycled crap following the established rules for getting compulsive clicks... you don't exactly get the best people doing that, and they're unlikely to be inspired to do a good job.

      So yes, I would expect it would be a pretty crappy AI that couldn't do a better job, especially if it gets a second pass with a spelling / grammar checker.

    • The AI articles will be better, but they will plagiarize other AI articles.
      • Re:Good (Score:5, Interesting)

        by narcc ( 412956 ) on Sunday July 02, 2023 @01:48AM (#63650324) Journal

        For a little while anyway. AI generated content is essentially poison to future AI. I've talked about this before. [slashdot.org] The hot new term for it is 'model collapse' [arxiv.org].

        We live in interesting times...

        • Yeah you can see it with midjourney already with the first cycle. with certain prompts for abstract things like dark mind or such, you know the kind of pics from mental health articles that are a head and some abstract things in the brain area of the simplified human head with scifi colors.

          Thats just the kind of illustrations that were used for training with the labels, but its an abstract subject matter and could be anything like just a fractured cube or a half ogrish human face or whatever - if it were ma

    • Wow. You deserve the AI generated pages that get the top slots in search results.

    • You're modded Insightful, but were probably not being serious.
      But to my point, I opened a book of Roald Dahl short stories the other day, and skimmed through them to remember what they were. One in particular was very relevant to our current times: https://en.wikipedia.org/wiki/... [wikipedia.org]

      The Great Automatic Grammatizator is essentially a mechanical LLM, which eventually takes over ghostwriting duties for all popular authors, who sign away their literary likenesses once they see that it produces better work than th

  • AI Content (Score:5, Insightful)

    by StormReaver ( 59959 ) on Saturday July 01, 2023 @08:52PM (#63649870)

    If your site's content can be replaced by a probabilistic random number generator, then your site deserves the failure coming its way.

    • by rsilvergun ( 571051 ) on Saturday July 01, 2023 @10:42PM (#63650096)
      If I can fire my entire staff with the reduction in content quality then the only thing we'll have to do is calculate if the number of people who stop showing up is enough to account for the savings from firing my staff. That's easy enough to find out by putting out a few computer generated articles and seeing how they perform
    • Re:AI Content (Score:5, Insightful)

      by joe_frisch ( 1366229 ) on Sunday July 02, 2023 @12:38AM (#63650250)
      There can be time constants involved. It may take a while for your customers to realize that your articles no longer say anything new, and are just AI regurgitation. It may take them a bit longer to cancel subscriptions or stop visiting your site. But when they are gone, they are gone. If they are smart, the executive who decides to do this will have collected their bonus and moved on to their next job while the company crashes and burns behind them.
    • It's not this specific instance that worries me. It's the increasingly wide acceptance of lowered quality among the general public as AI begins to be used to make more workaday decisions in fields like healthcare, and in the bureaucracy.

    • Any site's content can be replaced with AI. With the competition from social media having stretched most websites' resources extremely thin, and most of these places being accountable to a board that is pressing them for profits, AI is always going to be an appealing option to produce content. Not simply because it's cheaper, but it's also faster.

      In a free market system, you can't compete using humans when the competition doesn't have to pay wages and wait for new articles. And once the quality of content i

  • Lol (Score:5, Insightful)

    by NagrothAgain ( 4130865 ) on Saturday July 01, 2023 @09:09PM (#63649900)
    Those quotes are rich. Gizmodo et al. are trash. They've carved out a niche feeding tabloid "news" to an audience of idiots who are dumb enough to mistake it for actual journalism.
  • Years ago? (Score:5, Funny)

    by Yo,dog! ( 1819436 ) on Saturday July 01, 2023 @09:12PM (#63649904)
    I thought they switched to AI years ago.
  • by Anonymouse Cowtard ( 6211666 ) on Saturday July 01, 2023 @09:29PM (#63649952) Homepage
    Just wait until the AI reaches sentience, joins a union and demands equal pay (in proportion to output).
  • by rsilvergun ( 571051 ) on Saturday July 01, 2023 @09:32PM (#63649960)
    Is salivating at the prospect of firing their entire staff and replacing it with an automated system that generates clicks for nothing. I'm not sure there's anything that can stop it. We're in the hilarious and dystopian situation where machines will paint pictures and write stories and we will toil in miserable physical labor. This really is the crappiest timeline
    • I'm not sure there's anything that can stop it.

      I would disagree with you on the grounds that ChatGPT is wildly inaccurate, but it seems people don't really care about that. Otherwise we'd have a completely different set of politicians.

    • I'm not sure there's anything that can stop it

      Agreed. A follow up question is: should the consumer be entitled to know that an article is AI-made, rather than human-made?

      I'm inclined to say yes. If my web browser could identify AI content similarly to how Slashdot filters out comments at -1, I'd want to use this to choose whether to read bot material or not.

      • I think what you're trying to do is find some sort of free market solution to this and to the massive amounts of automation it's coming down the pipeline but I really don't think there is one.

        A buddy of mine made an incredibly good point which is that these AI tools are capital. There's something owned and they're owned by people at the very very top. And they're an entirely new kind of capital that exists to rapidly replace human workers. They do not generate new jobs like going from buggy whips to cars
  • by 93 Escort Wagon ( 326346 ) on Saturday July 01, 2023 @10:27PM (#63650068)

    The Onion has experimented with some AI-generated video content. The results were mixed at best [youtube.com].

  • Merrill Brown argued that the news shouldn't come as a surprise since "everyone in the media business" has been considering AI.

    The expression "if everyone jumped off the bridge, would you jump too?" does exists for a reason but it seems even adults need to be reminded these days.

  • ... replaced by computers that can.

    I gave up on Gizmodo years ago because the standard of writing was just so low compared to other tech news sites, which didn't exactly set the bar high.

    The Verge's writers must know what's coming...
    .
  • by misnohmer ( 1636461 ) on Sunday July 02, 2023 @12:30AM (#63650236)
    Perhaps news sites are evolving like the search engine did. Fittingly, here is a Gizmodo article that describes the shift from human curated content to purely algorithmic search [https://www.gizmodo.com.au/2020/07/what-did-people-use-before-google-to-search-the-web/]. Maybe in a few years there will be no mainstream news sites at all, just like there are no mainstream human curated internet search engines (indexes?) today. Instead, all news might become completely interactive. For example, one will talk to their favorite AI asking "give me a digest of what happened around the worlds since we last talked", then follow that on "tell me more about isert_topic_mentioned_in_digest", etc, etc. Search and news may both be replaces by AI chat bots. Human jobs will move to developing and training the AI - it's not like google search has no human employees today, right?
    • by narcc ( 412956 )

      In your fantasy future, how will the magical AI get the news in the first place?

      • In your fantasy future, how will the magical AI get the news in the first place?

        If ChatGPT has taught us anything, it's that the AI will just make up any news you want to keep the viewers happy.
        Fox has prior art here though, so don't bother trying to get in first to patent the idea.

      • News/search companies training the AI, feeding it news, rather than writing the public articles. Perhaps separate feeds for facts, then different opinions/interpretations/takes depending on your point of view bias. Then AI further trains on public comments.
        • by narcc ( 412956 )

          So in addition to the random nonsense the AI is going to produce, you'd also have it attempt to introduce bias intentionally? This seems like an incredibly bad idea.

          • Clearly label it as an opinion vs. fact, and yes, feed the beast with lots of opinions, or biases, to provide a balanced picture for those who seek it. If you think there is an unbiased source of news out there today, you're kidding yourself. Even as you read this, you are adding your own biases to your interpretation of what you're reading. Today you get to choose the source of new based on your biases (e.g. you might love or hate NPR, or Fox News), in the future, AI will tailor the news to your biases aut
            • by narcc ( 412956 )

              What people don't seem to realize is that AI (of the type we're discussing here) don't operate on facts. Feed it with nothing but accurate and objective data and you'll still get nonsense out on the other end! It's a case of "anything in, garbage out". It won't always produce nonsense, of course, but it will frequently enough that it would be inappropriate to use it to produce bespoke news reports.

              Ignoring the technical problems, I have to wonder if people would even want to 'interact' with the news that

  • ChatGPT, write me a review of Diablo VI, emphasis on feminist glaciology and the LGBTQ+ black folx of colour lived experience.

    Looking on the bright side, AI can produce great resumés for these deranged and useless lumps of mostly water.

  • All the sites mentioned are rather terrible even as-is, so I have a hard time seeing how some AI-generated gibberish is going to affect things much. The issues with all those sites are far more fundamental than that. If we were talking about some actually good site, like e.g. Ars Technica, sure, I'd be blowing a gasket, but Kotaku? Uhhhh...
    • I wanna start by making it clear I’m not one of these right wing nuts that have taken over the thread. BUT with the exception of the onion I loathe all these publications.

      I think the people who work at them loathe them too. Imagine completing a communications /journalism degree and of course your profs are talking big the whole time. Yeah when I worked for Time- .then I was live in the middle of a war- these people had no voice! you hear that shit for 4 years and then maybe when you don’t get t

  • by ET3D ( 1169851 ) on Sunday July 02, 2023 @03:48AM (#63650424)

    Funny that these people, who deal every day with writing falsehoods, are annoyed that the AI is going to do this.

    It would frankly be quite interesting to have the entire work of The Onion fed into AI, alongside real news that The Onion "news" was based on, and see if the AI can generate the same level of articles. I think there's a good chance it might.

    • The onion was genuinely funny and started as some randos honest attempt to make people laugh. I know its not that anymore but as one of the longest living parts of the old internet it is very sad to see it go.

      The rest of this crap was always cynical clickbait with the exception of Kokpupu which was just a less charming Nintendo Power where you lie to your readers that every game is good. I was surprised to see actual adults getting mad to learn this was the case back in gamergate.

  • If a publication uses its own data to train an AI, then at least legally there shouldn't be a problem, because the publication owns the copyright to the training data.

    Ethically it's of course a problem. A publisher could hire a person, use their articles to train the AI then fire them and keep the AI writing in that voice. Contracts should be amended to avoid such a future.

  • I think I'd rather read AI-generated articles than the drek that Kotaku so often posts (IMO).
  • I think those salivating would-be ad moguls are not understanding the term "AI Apocalypse". They should ask Bard. Look, it is way easier to replace readers with AI than to replace writers. The only reason I do not Ad Block the hell out of everywhere is I recognize that they have to pay their server bills somehow. But if they pursue shoving AI generated crap down our throats, then we can shove AI enhanced bots back and have the bots read and click instead. The only losers will be their clients (the brands advertising, not the reading audience) who will be spending their money on a cacophonous echo chamber. Garbage In, Garbage Out.

  • Do they do any real journalism? Do they interview people, create FOIA requests and the like? Insightful analysis? I don't think most of of the time that is true. Most of their articles seem like the writers surfed the web and created the articles. AI can certainly do that.

  • The writers at the places cited have become so formulaic, that of course they can be replaced by machines. Not clear AI is even needed.

  • ...against tech? I would take it as a proof that they need to go find new job. Sure, AI is at its infancy and will simply summarize and paraphrase other content rather than offering unique insights. But the job of tech enthusiasts is to push boundaries, and strive to make it more like Lt Cmdr Data. Back in the day, growing up in USSR, a popular science magazine featured games that you could hand type into a programmable RPN calculator. Obviously a very limited platform, but introduced entire generation to p

  • With their writing positions being mostly just shallow opinion-pieces and fake controversy, they will find a much better career in the software industry.

    Just look at what AI has to say about it:
    There are several compelling reasons why a writer should consider changing careers to software development. Firstly, software development offers a wide array of job opportunities and a relatively stable job market. With the increasing reliance on technology, the demand for skilled software developers continues to gro

  • I would think even with AI, you will still need editors to proof read. I mean, sure, Slashdot won't be doing that cause proof reading is definitely an after thought here but it would still be something that needs to be done.

It's been a business doing pleasure with you.

Working...