Over the last year, odds are good that you’ve seen what’s been dubbed “AI slop”—unhinged, nonsensical “art” generated by artificial intelligence tools. Maybe you’ve seen a bizarre cinematic animated mini-movie on Facebook, surreal pseudo-photographs on Instagram, or propagandistic images on what was once known as Twitter, now X. After seeing enough of this, a realization dawned on Jacob Silverman, a journalist in New York who covers technology and politics: if it’s machines making this art, and bots who are showering them with likes, where do humans fit in?
The answer is that actual living people are being squeezed out of what Silverman has, in a recent Financial Times article, deemed the “hostile internet“. Elon Musk’s X will sell advertisements, and authority, to absolutely anyone; AI-powered chatbots are worryingly easy to manipulate; and it has never been easier for people suffering from mental illness to find positive reinforcement of their ideas, both from distant humans and AI. None of this is to the betterment of humanity.
Silverman joins Phoebe Maltz Bovy on The Jewish Angle to discuss these trends of digital devolution, and how we can navigate these murky waters on a sinking ship.
Transcript
Phoebe Maltz Bovy: Hi, this is Phoebe Maltz Bovy, and you’re listening to The Jewish Angle, a podcast from The CJN. The Internet these days can seem kind of awful. And by awful, I don’t just mean the kind of hateful pylons of yore. I don’t mean people yelling at each other. I don’t mean even people holding necessarily extremist views, although that’s all been unpleasant all along. But it can just seem kind of nonsensical, awful in this nonsensical way. I remember first noticing this on Twitter once Twitter started getting ads nonstop for bras for elderly women, and there were these various elderly women in different bras. It didn’t seem to be. At first, I thought, are they calling me elderly? Have I gained some decades without being aware of this? But no, no, this is just everybody was getting these ads for these elderly women bras. I am now nostalgic for the era of those ads because I don’t even know what I’m looking at now. And here to help me figure this out is Jacob Silverman. Jacob is a technology journalist in New York and co-author of Easy Money: Cryptocurrency, Casino Capitalism, and the Golden Age of Fraud, a New York Times bestseller. His new book, Gilded Elon Musk and the Radicalization of Silicon Valley, will be out in October. I’ve asked him on because I just read his new Financial Times article, “Welcome to Slop: How the Hostile Internet is Driving Us Crazy.” It’s really on my own social media making the rounds, and rightly so because it’s got the best description I’ve seen of what online life has become. And I just do one quote from it and then I’m going to introduce Jacob. So it’s neither mass broadcast nor targeted communication. The posts, these are these promoted posts on Twitter, now X, landed in some netherworld of inscrutability, their meaning known only to their promoter and maybe not even to them. Jacob, welcome to The Jewish Angle.
Jacob Silverman: Thank you. Glad to be here.
Phoebe Maltz Bovy: So my first question for you, or really more on behalf of any listeners who have not yet read your piece, which they should do. What is the hostile Internet?
Jacob Silverman: So this is my attempt to humbly coin a neologism, I suppose, but I am talking about the daily experience of using the Internet, whether you’re talking on any device really, but of conducting, talking to people, doing work, going through your apps. You know, the whole kind of daily experience has become full of obstacles and very adversarial as an experience, and very frustrating and confusing. I don’t just mean because I’m not talking about people being technically inept or lacking any kind of media literacy, but it seems like every time we log on, there is kind of something standing in our way or an algorithm trying to guide us to some other desired activity. One line I have in the essay is, I say that today’s systems aren’t really optimised for human flourishing or for what’s best for us. They’re optimised for, according to metrics that, of course, profit the platform owners. This manifests itself in a number of ways. But one of them is these kinds of inscrutable ads and various kinds of what’s called AI slop content that’s just filling social media feeds now to varying degrees.
Phoebe Maltz Bovy: So I have many questions about various angles of this, but first off, just which platforms are we talking about here? So is this a Twitter thing specifically? Is it also Facebook when you Google things, what? What is it? Where is it happening?
Jacob Silverman: So I started, or kind of centre it in X right now, because that’s where I started seeing a lot of these ads personally that were coming to my feed that, like the line you quote, they, you know, they weren’t directed towards a specific audience, it seemed like, but they also weren’t an advertisement. They weren’t addressed to a mass audience because no one could really figure out what was going on. Like one of them was a poll about that said, “The most sentimental gun you own,” like inherited it or bought it. It’s like that, you know, it’s just sort of a non sequitur. It’s baffling. And then I centre the article a bit or open it with this woman in Michigan who, you know, we’re talking a bit about, about people who might be mentally disturbed in one way or another. But. And there are a lot of people like this woman in Michigan who seem to be buying ads on X and other social media to promote their posts, where they are presenting their sort of skewed or schizophrenic view of the world. In this case, this is a woman who’s kind of live streaming her activities from a motel, one of these chain hotels in Michigan. I started watching some of her videos and it’s clear that she thinks she’s uncovered something that Elon Musk and Donald Trump need to know, and she wants them to come visit her at the hotel. Meanwhile, she posts these videos where police are visiting her, saying in a rather gentle style for American police, you know, please just kind of tone down your activities. Don’t film in the lobby. There are complaints, things like that.
Phoebe Maltz Bovy: And when you say. Just to interrupt, when you say schizophrenic, you don’t mean quite the. You’re not armchair diagnosing this woman with a specific mental illness. Because I think from your article, it’s clear what you mean by it, but our listeners might not understand.
Jacob Silverman: I don’t want to seem like I’m providing a diagnostic here, really. It’s more. We’re kind of more talking about behaviours that would strike one as mad. I even talk about it in the sense of we’re all talking to disembodied voices now or hearing them through our phones as we’re having conversations walking down the street. It’s adding to sort of the growing incoherence that I’m talking about. You know, I don’t really know what’s wrong with this person, but the way her expression and these ads come through is schizophrenic in the sense of lacking meaning and connection to reality and seeming to come from some other place. That’s the kind of effect, that alienation, that dislocating effect that I see happening on a lot of platforms. On somewhere like Facebook, it’s happening more with the AI slop images that have gone around. One of the popular ones from last year was called Shrimp Jesus. It was like a Jesus made out of shrimp.
Phoebe Maltz Bovy: I don’t know, it’s like, how did I miss that? That’s very non-kosher.
Jacob Silverman: You can kind of imagine now that we’re somewhat versed in AI apps, like what was the problem prompt that made this image? But again, these are like, you know, art without reference. Things that are just sort of artificially generated and designed to fill a content-shaped hole and that make the kind of real estate or landscape of your online world confusing and incoherent and non-communicative, you know, not communicative.
Phoebe Maltz Bovy: Is there a way of. I’m wondering if there’s a way. So. And I’m thinking of my own use of the Internet and how I used to be on Twitter constantly or X now, and I went off of it. Not off of it like in a dramatic flounce from it. Like I still use it as effectively an address book. But what put me off wasn’t the extremism. Not that I liked the extremism, but it wasn’t new to me, the extremism. I worked at the Forward during the first Trump administration, and I’m used to the Nazi trolls. I don’t like them, but they didn’t seem new. What did seem new were these incoherent ads. However, I’ve hopped over to BlueSky, which doesn’t seem to have that issue. It may have other issues. Is the answer to just use BlueSky?
Jacob Silverman: Well, I think for some of these issues that we’re talking about, X has exposed itself in a particular way. In that Musk very vividly told advertisers to go F themselves, and as a result, they’ve opened up the ad network on X to anyone who wants to pay to promote a post. They push all these notifications and inducements towards users, so you get the kind of devoted but often delusional user base buying in there. I think the answer is in that case perhaps to use BlueSky because these smaller spaces, perhaps more curated, moderate, or community-based in some way, might be the kinds of digital spaces that preserve the values that we’re looking for. This would avoid the influx of slop and incoherence we’re seeing elsewhere when things are of a reasonable size.
Phoebe Maltz Bovy: Yeah, I mean, BlueSky—it’s been. I wrote something not rave review of it and then got blocked by half of it for that. Even though I am not MAGA, I got put on block lists for MAGA and things like this, even though I was very much Vote Harris as much as you can from Canada. Another line from your piece that I really liked was, “The result is less a broadening of the public school square than its pollution.” I really liked that because it stuck with me. Even if you approach social media as someone who wants to see different perspectives—not necessarily in a “debate me, bro” sort of way—there’s this absurdity in wanting to strengthen your own arguments. It seems like the alternatives right now are sort of slop or bubbles, possibly.
Jacob Silverman: I think on the social media front, yes. Maybe this is also an argument for returning to primary sources, to engaging with written media, whether you’re reading newspapers and magazines or participating in the healthy activity of reading books and putting your phone away. It says something about the failure of these spaces of public discourse, which is why I mentioned this idea of pollution. Free speech is of primary importance, but when talking about it in the context of social media, there are other related or sometimes overlapping values. What we’ve had, especially on some of these less moderated networks, is a cacophony. If you have a cacophony or pollution of the space, it makes it less free for others and a less communicative place where people can actually have discussions and be heard.
Phoebe Maltz Bovy: So to ask about the question of the spaces not being what’s best for people but sort of serving broader aims of tech CEOs or whoever, in a simple sense, what you’re describing is just capitalism. You make an analogy between the hostile Internet and hostile architecture where seating is gone in case somebody homeless might want to stop by, but you can buy stuff everywhere. I get that, but also people have always been trying to sell you something, you know what I mean? That itself is not new. I mean, read Zola’s “The Ladies’ Paradise” and the department store. People have been trying to sell stuff, not you personally, but all of us, for a long time. So, are we looking at something new? Is this a new stage of capitalism, or is it something different?
Jacob Silverman: I wouldn’t say it’s something different. It might be a new phase of digital capitalism or surveillance capitalism as it’s been called. Over time, these surveillance-based business models of data collection and ad tech seem to degrade the quality of the platform or space. Cory Doctorow has talked about this idea called “enshittification.” One of the problems with these platforms is the same one they’ve had for 10 or 15 years. They don’t really have other business models and still operate according to the capitalist extractivist logic that underwrites a lot of the consumer Internet of data collection and advertising, monetizing that data. Until that fundamental problem is solved—so maybe I’m talking about capitalism itself—we may not get very far but will continue to find new permutations of it that feel more exploitative.
Phoebe Maltz Bovy: One of the most fascinating things in your piece is you describe the Internet as just sort of playing out in this zombie way where the entities interacting are not human.
Jacob Silverman: Yeah.
Phoebe Maltz Bovy: Do you want to talk a little bit about that? I found it fascinating. I would not be able to summarize and explain it.
Jacob Silverman: The theory is called dead Internet theory. There’s a formal literature to it now, but basically, it’s the idea that there are a lot of bots online and a lot of automated activity. In a literal sense, this is true because you have computers communicating with each other from banks and institutions, just running infrastructure. A lot of data and stuff gets transmitted without people talking to one another. Now, this is happening on the human layer of the Internet. When you’re on social media, you’re as likely to be responded to by a bot or some non-human entity or software program as anything else. Beyond that, you have bots interacting with bots. Forms of this have existed for a while; there have been allegations in digital advertising of tremendous fraud where ads are shown to no one, appear far down a page, or have lawsuits about this. It’s a multibillion-dollar issue in the online ad industry. The new version is that bots watch the ads and click on them. As the layer of the Internet we occupy becomes mingled with these automated agents, it feels alienating and adds to the sense of hostility, confusion, and inhumaneness.
Phoebe Maltz Bovy: Yeah, so I was just gonna say, everything online seems so stupid, yet it also makes me feel stupid. I’ll be in an argument on social media and start wondering who or what I’m even arguing with.
Jacob Silverman: Yeah.
Phoebe Maltz Bovy: I’ve felt, and talked about this on another podcast, like I’m arguing with a cartoon character. Maybe that’s another real person, but I don’t know that. This especially happens when I’m arguing with someone with a pseudonym or cartoon avatar. But they are so much more righteously whatever it is than I am. They’re so brave in how they are using even harsher anti-Trump language. It’s like, but who are you? What are you? Like, I don’t know, what are you risking here?
Jacob Silverman: Yeah, what’s the point?
Phoebe Maltz Bovy: Yeah, but I mean it’s just like, I guess I would say that life online feels a bit like when you’re getting a scammy phone call, except that it’s constant and you need to be so savvy to know what’s a scam. And also, like with the AI stuff, I feel like I will recognize AI in the arenas that I’m familiar with. Like with writing, it looks obvious to me, but like I’m not as musically minded. And like with music, I wouldn’t necessarily know if a jingle is AI. No, I know that I wouldn’t know because this has been tested.
Jacob Silverman: Yeah. And I think that that kind of scam prevention mindset, sort of having your head on a swivel and trying to be more alert and aware and self-questioning, is one that more people are feeling or adopting online and probably should be. Because we’ve probably all talked to elderly parents or relatives about like, no, don’t answer that message or put down the phone. But it seems like we are constantly encountering opportunities to be grifted or scammed or manipulated online. Whether it is outright, you know, someone trying to steal your money, or a fake item being sent to you, or clicking on an ad that you don’t really want to click on but seems like the article that you need, or wasting your time engaging with someone on social media who may or may not be real or whose intention may be, “Hey, I want to waste this journalist’s time.” You know, there’s that too. That resonates with me, what you’re talking about, the sense that you kind of have to approach a lot of your time with a mindset of caution now, I think because otherwise you’ll end up either with your pocket picked or in the wilderness arguing with someone for an hour.
Phoebe Maltz Bovy: So these kind of two questions, how much of what you’re talking about is geographically specific or US specific? And also, insofar as you can answer this, how much of this is different, maybe worse in Canada?
Jacob Silverman: Sure, this is a pretty global phenomenon. Actually, some of the things we’re talking about are globalized in the sense that, as I quote the journalist Max Reed, who has written some good stuff about this, about the AI slop invasion specifically, there are entrepreneurial forces behind this. I mean, you’re not just seeing AI images in your feed because it’s funny or something like that. It’s because people, often in the global south, you know, in Nigeria or Southeast Asia somewhere, are creating this stuff and operating a lot of accounts and leveraging the advertising and reward systems of these platforms to get paid. So they are pumping this stuff out on various social networks and promoting it, then getting paid for the engagement. And of course, there’s no real human in the loop on the platform side, whether it’s Meta or X or whatever else. The platform doesn’t really know or doesn’t seem tuned to care that this guy is not quite a scammer but is making money by churning out junk. And it reflects this broader principle that anything you could call content will do.
Phoebe Maltz Bovy: Except but not news, right? So it seems like you have this issue. Oh yeah, yeah.
Jacob Silverman: I’m sorry about that. I should be specific about that. There are two things about that. One is that when Facebook has removed or been removed news material with these national disputes it’s had, like when they did this in Australia, they kind of say, well, news isn’t a big part of people’s feeds anyway, which is true in a percentage sense. But it of course just leaves more room, more of a content-shaped hole. And what do you fill it with? More slop or junk. So I think even when Facebook or one of these big platforms is operating, quote-unquote, as it should, you still have legitimate news or more desired and legitimate material mixing with the slop and always being infiltrated by it. So I would say right now probably, though with news not being really available on Facebook, I would think it’s worse. I haven’t been able to experience it much myself lately.
Phoebe Maltz Bovy: There’s a quote from your piece where you write, can we find a way back to an Internet that puts people in lucid conversation with one another? And I’m thinking about just how interesting it is to see how easy it can be now to feel nostalgic for things that actually at the time seemed awful, like clickbait or exploitative personal essays where the authors were paid peanuts, if anything, for something that would stick with them forever. It now feels kind of quaint to think about that as having been problems when you think about what we’re looking at now. So are we looking at something that’s just an acceleration of that, or is it just different in nature?
Jacob Silverman: Yeah, I think it’s sort of a furtherance or intensification of that trend where any expression or what we casually call content now, but was probably sort of a dirty word to us a few years ago — at least to me, it felt a little uncomfortable to say — calling everything content is of decreasing value. One sort of dynamic in this or one catalyst is that the AI companies, which is to say the big tech companies, now have to consume as much data and content as possible to train their AI models and try to develop them further. Some of this is done through licensing deals and some through basically intellectual property theft. And that has contributed tremendously, I think, to the further devaluation of what we call content, pushing us further to this place where any piece of media will do as long as there’s a human looking at it or someone we can claim as a human according to the metrics.
Phoebe Maltz Bovy: This may be a little out there, but one more thing that is maybe out there, it’s just like, I’m wondering if there’s any risk in romanticizing earlier eras tech-wise, because I do remember the era before slop. But it was also one where people who were definitely real and often extremely smart and accomplished bullied one another relentlessly. I am not of the right-wing anti-woke persuasion. I have, however, written about callout culture, as it was called at the time. And it was very intense. Some people did lose jobs over it. You know, the Internet is obviously worse now, like in the sort of ‘Inshinification’ sense of worse. But it’s also like what I find now is, it’s just much easier to tune it out. I don’t remember the last time that I felt emotionally invested in anything online. So what I’m wondering in a roundabout way is, is it possible that things just get so ridiculous that everyone who’s not in a sort of tinfoil hat brigade logs off and that’s where this all goes.
Jacob Silverman: Well, I do think the fact that people have really taken to group chats in recent years is a natural thing, of course. I think we’d want to talk to our friends like that, but I think it’s also reflective of the fact that these broader open platforms have become hostile or unpleasant places. And I think that will continue because, you know, that’s a core part of communication to me. And I think also, in terms of the nostalgia question, I am wary of nostalgia kind of broadly, but also kind of Internet nostalgia. I’ve written about a little bit that I think it can be misplaced sometimes.
Jacob Silverman: So when I talk about going back to a place of lucid conversation, I think it’s more important to talk about values or the kind of environment we might want to create. It’s not like we need to recreate the Internet of 2007 or a particular platform that I loved and that spoke to me during my formative years of going online, thinking, and talking to people. That’s where we sometimes end up—it becomes about individual taste or even nostalgic memory, not really about reimagining things or discussing values that do matter to us. One of the things we’re getting at is the value of closer communication—whether it’s in group chats or smaller networks. I am a little nostalgic about message boards dedicated to particular topics. That’s not for everyone. It can be more fun as a teenager or college student when you’re on your favorite band’s message board. But it indicates that smaller communities might represent what we’re looking for, at least to some extent.
Phoebe Maltz Bovy: That’s interesting. Yeah, I’ve learned from colleagues at the Canadian Jewish News that more Observant Jews use WhatsApp. That’s what they use, and therefore they avoid that ear-dropped, airdropped earwax, whatever it is, let’s leave that in. So, Jacob, thank you so much for coming on. Please tell us where and when Gilded Rage is available? And where can you be found on these dangerous Internets?
Jacob Silverman: Yeah, I am at jacobsilverman.com. I am on X as Silverman Jacob, and my next book is available for pre-order on most online retailers. It’s called Gilded: Elon Musk and the Radicalization of Silicon Valley. It’s about the rise of the tech right and some of these issues we were talking about today. So, thank you.
Show Notes
Credits
- Host: Phoebe Maltz Bovy
- Producer and editor: Michael Fraiman
- Music: “Gypsy Waltz” by Frank Freeman, licensed from the Independent Music Licensing Collective
Support our show
- Subscribe to The CJN newsletter
- Donate to The CJN (+ get a charitable tax receipt)
- Subscribe to The Jewish Angle
