Newspapers Are Recommending AI-Hallucinated Novels
The GenAI Slop Flood Has Overflown the Internet Container
Over the weekend, the Chicago Sun-Times—a storied and award-winning newspaper and longtime home of Roger Ebert—published a summer reading list. Almost all the books were fake. There is no Nightshade Market by Min Jin Lee, Boiling Point by Rebecca Makkai, The Last Algorithm by Andy Weir, or The Rainmakers by Percival Everett, among other invented titles. The article was not only generated by ChatGPT (or similar program), but clearly unedited. No one at the Chicago Sun-Times even bothered a cursory check. And not only the Sun-Times. The article, along with other seemingly AI-generated pieces, were syndicated in multiple newspapers across the country including the Philadelphia Inquirer.
Such AI slop is already ubiquitous online. Social media, clickbait websites, and the self-pub ecosystem are overrun with AI nonsense pumped out without even a glance. It costs virtually nothing to flood the web with slop and hope for a few sales or web ad payments. But it seems novel, and extra worrying, that such slop is now appearing in a printed newspaper. For a while one could hope that—even if the internet was ruined by spam, slop, and scams—physical media would be somewhat immune. Apparently not.
Some of these fake titles appear with fake summaries in Google searches, although it is unclear if Google is duplicating the paper’s error or vice versa.
For its part, the Chicago Sun-Times is “looking into how this made it into print.”
[Editing to add: the following statement from the Chicago Sun-Times Guild]
I have to stress again that you cannot have published a piece like this if you actually edited it. Or, hell, at least did a quick Google check. The “author”—or I guess “prompter”—of the list confirmed as much to 404 Media: “I do use AI for background at times but always check out the material first. This time, I did not and I can't believe I missed it because it's so obvious. No excuses.” It’s obvious this was more than just background though. Even the real titles have poorly written summaries. But this is how GenAI use was always going to go. Even if you tell yourself you will use it responsibly and that it will “enhance” your work, you quickly give up. Why do any work when you can get away with doing none, after all?
Despite what techno-optimists have claimed, AI use is nearly always lazy. Yes, exceptions exist. But they are exceptions. The point of AI, for most people in most circumstances, is to do less work. “Less” soon becomes “as little as possible.” And “as little as possible” ideally becomes “none.”
This is true even of people’s “creative” work, where authors are already being caught leaving ChatGPT prompts in their (self-)published novels. This means they didn’t even glance at—much less edit or revise—the AI outputs they published as their own work.
This is something I’ve noticed as an undergrad professor. It isn’t just that the use of AI to skip work has become ubiquitous. It has also become lazier. Students who in 2023 would edit a ChatGPT output to try and avoid detection will now turn in “work” that doesn’t even remotely follow the assignment or that includes text from the chatbot conversation showing they didn’t even glance at the document before submitting. Not that I should single out students. The same issue is seen in teachers. A recent New York Times article cites a Northeastern University student who noticed her (yes, business) professor using AI: “She reviewed her professor’s slide presentations and discovered other telltale signs of A.I.: distorted text, photos of office workers with extraneous body parts and egregious misspellings.” There are other funny examples in the article, including a professor’s comments on a paper accidentally including “a back-and-forth with ChatGPT” asking for “really nice feedback.”
What was even more telling to me, though, were the professors saying they used AI to deal with “overwhelming workloads.” I believe them! But this also points to where AI is going. The correct response to AI for academia would have been smaller class sizes and more full-time faculty. Much of academia now is run by overworked adjuncts with insane student loads. If you are teaching 150 students at 5k a class, why wouldn’t you lean on AI to reduce the workload? And if you are a student in a gigantic class with a professor who can’t possibly monitor much less help every student, why wouldn’t you turn in AI work and spend your time on other things? But academia seems unlikely to do the educationally enriching yet expensive move of hiring more professors and shrinking classes. Instead, they’ll let professors use AI to help with the overwhelming workloads… then increase those workloads since “now you have AI to help you.”
Academia has been the focus of much of the AI discourse, but I think it is clear this is the near future of nearly every aspect of life. We’ll see the same thing in every workplace that is able to implement AI, no matter how unhelpful the programs are. Hell, even getting a job will require jumping through hoops of AI slop. Slate had a piece this week about companies using glitching AI programs to interview job applicants:
“I’m so excited to talk to you and get to know more about you,” the bot says, before immediately falling into a loop of gibberish. “For our first question, let’s circle back. Tell me about a time when—when—when—when—let’s. Let’s—let’s circle back. Tell me about a time when—when—when—when—let’s.”
One interviewee is quoted in the article as saying “it was very disrespectful and a waste of time.” This is what I thought when ChatGPT first came out: this will waste everyone’s time. More time wasted dealing with customer service chatbots that can’t help you. More time wasted wading through slop on social media. More time wasted on AI-generated emails you don’t want to read, on trying to find accurate news, or even on just finding a good book to read that actually exists.
Well, at least for that I can log off and pull an old book off the shelf. Maybe I’ll do that now.
For a novel published without (I promise) a single bit of AI use, check out my new novel Metallic Realms. It is out in stores and getting great reviews: “Brilliant” (Esquire), “riveting” (Publishers Weekly), “hilariously clever” (Elle), “I left this book feeling … assured of the future of storytelling” (The Masters Review), and “Just plain wonderful” (Booklist). You can also read an excerpt of the novel in the excellent Substack magazine The Metropolitan Review.
.
Speaking also as a professor to undergrads:
I have observed the untempered CONTEMPT of some students for the others they observe using chatGPT in class. They come and tell on their peers after class, SEETHING, or they turn in assignments that say things like (actual quote from in-class writing assignment): "Or, for good example, be like, to give a random name, [ name of actual student he sits next to ] and use chatGPT to give you the answer since you desire entertainment so much that your brain has shut down." (I already knew there was a problem with Student-Redacted because he was unable to answer simple classroom conversation questions orally. But this seething hot anger and contempt for their cheater-peers is the thing that gives me hope for the future: because other than the people using these tools, everyone else finds them insulting to their humanity, even if they're not even the ones who have to read the slop (and especially if they are).
It does seem like anger has become the appropriate (normal) reaction: to feel insulted, dehumanized, to feel that what the bot-text-monger is doing is degrading, perhaps to civilization or humanity itself, but DEFINITELY to the human being whose eyeballs were tricked into reading slop.
First of all: we are doomed! But then silver lining is in understanding what is happening so I have to recommend the excellent "The Eye of the Master" by Matteo Pasquinelli on the topic of AI, its history specifically under the aspect of division of labor. Must read. And lastly I will take Michel's advice and grab an old book from the shelf: "Der Bauernkrieg" by Friedrich Engels (in English: "The Peasant War in Germany"), originally published in 1850 my edition is from 1946 in pristine condition. Let's go in the park, wander around, reading, like the good folks in "Fahrenheit 451". VIVA LA REVOLUTION !!!!!