Michael Calore: So publishing is definitely at the top of the list of industries that have been worried about AI plagiarizing original work, and we should all know because we’re all in the publishing industry. But then there’s the content that is the opposite of thoughtful, human-made work, and that is AI slop. The term explains itself when you say it out loud, but let’s quickly talk about what AI slop is and why it seems to be everywhere.
Lauren Goode: I can take this one, but also, I do want to toss it back to Kate, because Kate, you are the queen of AI slop, and I don’t mean that you generate it. I don’t mean that it’s part of your personal content creation vector or whatever we’re calling it, but you’ve written a lot about it. AI slop is just low-quality, shoddy AI content that is appearing online. It is proliferating our feeds. It’s often on social media, but it’s not just on social media. It is now being passed off as legitimate, quote-unquote, “journalism”. For example, last month, the Chicago Sun-Times and the Philadelphia Inquirer had both published these special sections recommending summer reading lists, and the list included a bunch of made-up books by real authors, and these names and titles were just thrown together at random. Slop isn’t just made-up stuff though. I think it’s got a certain aesthetic. It’s part of this growing trend of the enshittification of the internet, which of course Cory Doctorow wrote about for Wire.com a few years ago and now I’ts just the term we use. It feels like spam, and sometimes it’s easily recognizable and sometimes it’s just not.
Katie Drummond: So you mean the videos I see on TikTok of Donald Trump and Jesus Christ walking on the beach are not real?
Lauren Goode: No, those are real.
Katie Drummond: Oh, OK. That happens.
Lauren Goode: Those really happened.
Katie Drummond: Oh, OK. Because I’ve been faving all of them, because I want to see more. So those are AI. Got it. OK.
Lauren Goode: Yes, exactly. Same with JD Vance breakdancing with Pope Leo, those are real.
Katie Drummond: Oh, I have … Yes, of course.
Lauren Goode: Yeah. Hasn’t killed him yet.
Michael Calore: A lot of these examples are funny or fun, but then there are ones that are more serious. There was AI slop coming out of current events in the Mideast recently, right?
Katie Drummond: Oh, of course. Yeah.
Michael Calore: And politicians and world leaders will retweet these things, even knowing that they’re fake, just because it appeals to their sensibility and it helps them spread the message they want to spread.
Katie Drummond: Oh, I make jokes when I’m stressed out and uncomfortable, and I would say it is incredibly uncomfortable and stressful. I think you would all agree with me being a journalist right now. Try being the editor in chief, let me tell you. And actually watching AI slop proliferate across the internet, across all these platforms, sometimes be mistaken for factual information by consumers at the same time as we are in this very existential moment for news and media. Yet again, we are in an existential moment for news and media, in many ways because of AI, because of the way Google is changing their search, because of other ways that AI is changing how people access information. Publishers once again are essentially in the crosshairs of all of that, and to add insult to injury, you then open TikTok and Jesus and Donald Trump are fishing, and it’s just like it’s everywhere. It’s like it’s surrounding you if you are a journalist because you were experiencing the slop itself. You’re seeing what it’s doing to the information landscape online, and then you’re banging your head against a brick wall because Google did this, that or the other thing with AI overviews, and all of a sudden I’m inventing numbers. I genuinely am inventing numbers, but all of a sudden, your search traffic is down 50%, and that has existential ramifications for publishers. There’s also this weird thing happening that has caught my attention, and Kate, you’ve reported on this, which is where AI generated content is actually like a feature for some websites and actually works really well for them. So WIRED found that over 54% of longer English language posts on LinkedIn, everybody’s favorite social network, are likely AI generated. Now, LinkedIn have said that they monitor posts to identify low quality and repetitive content, but AI is probably really good at LinkedIn because generic, bland writing is kind of what LinkedIn thrives on. I think that that’s interesting. It’s not necessarily a good thing, but it’s just another indication of how pervasive generative AI has become online.
Read the full article here
