No disagreement. We might come down on this differently, but that’s why the challenge I’m trying to highlight here is that, as readers, we almost have to be agnostic about its use. At this point, that’s what I’ve chosen to do. It’s already all over the place, whether I like it or not. But when I ask myself what I object to as a reader it’s bad writing—AI or otherwise. The Jeff Goins post I linked to above captures my feeling of revulsion and AI-inflected prose. When I start seeing those tics and tells I get irritated. But not because it’s AI. Rather, it’s because those tics and tells reflect lousy writing. More than reflect bad writing, they are bad writing. The product betrays itself.
"I personally don’t care how enthusiastic Tokarczuk or Qudan are about their LLMs, or even how they use them. I care about their artistic judgment displayed on the page. Is it any good?"
With respect, I think I disagree. This stance fails to grapple with the fact that writing has long had a strong social or communal character. In my own field, I see the extraordinary literary production of Greek writers in the 5th century B.C. and Roman writers in the Age of Augustus. One suspects that without Herodotus, no Thucydides; without Sophocles, no Euripides; without Plato, no Aristotle; with Vergil, no Ovid. So we have to ask: what is supportive of a vibrant community of writers who are continually influencing, inspiring, and challenging one another to higher and higher levels of achievement? Put differently, it can't be just about the product; it has to also be about the people behind the works. A writer who looks to her preferred AI—not to her friends and rivals who are nearby likewise sweating over their manuscripts—for creative sustenance is not giving anything back. In effect, her relationship to her cohort is parasitical. She wants to benefit from what that community generates, e.g., a kind of nursery for creativity and a reading public eager to buy books, without contributing to it. She thinks, "I have my AI. What do I need them for?"
That’s a valid point. But I don’t think it’s binary. Writers who use AI can still be—probably will want to be—in community with other writers. Humans need humans; to one degree or another, we like company. And yet plenty of writers have been reclusive types who self-isolate. I don’t think it’s either/or. It’s also possible something new will emerge that could prove good or bad—but we’ll have to wait to see what the outcome is.
I hear you. But one obvious follow-up question is this: when you do participate in our community, what do you have to offer the rest of us? What skills and insights *acquired by blood, sweat, and tears* can we learn from?
"I used AI merely to organize my thoughts."
"Merely? Okay, maybe someone else here did it the old-fashioned way and can help me with this."
Isn’t that self-correcting? If an author has little to contribute, they stop getting invitations or feel unwelcome in the group. They have to bring up their game to participate—which is true whether they use AI or not. No one likes working with people that don’t pull their weight.
At the same time, there may be new communities that form—probably will be since humans are incurably social—where writers who use AI find common cause. But even there, if someone comes in saying, “I used AI to merely organize my thoughts,” they’ll be shown the door too. No communities work without reasonable levels of contribution from the participants. Freeloaders aren’t welcome long.
So your revised thesis might be something like this: "As a reader, I care about artistic judgment displayed on the page. As a cultural critic of all things literary, I care deeply whether and how that discernment is acquired and would caution any upcoming writer that there are no shortcuts to becoming the next Joseph Conrad or Virginia Woolf."
I would endorse that. I don’t think there are shortcuts to judgment. That only comes from experience. Which does expose a real, potential threat with people using AI without the benefit of “learning the hard way.” For what we value right now, that’s the only way to learn it. It’s possible that’ll change—cultures evolve in ways current people can’t imagine.
This seems like a helpful way of thinking about it--including for academic scholarship where these debates are happening thick and fast, and where journals are currently trying to legislate if and how AI can be used (talk about closing the barn doors after the horses have bolted...). If your writing doesn't hold up to scrutiny because your claims are false, your metaphors are nonsensical, or your writing is the literary equivalent of Instagram face, then it's flawed, whether you got there with or without the help of AI, and moreover that's on *you* as the author (er, the generic 'you' obviously, not you as the author of this post).
Increasingly, I think that’s the only way to handle it. The process is very important for a writer, but the product is what the reader encounters. If the writer undermines their own process by overreliance on AI (the deskilling fear, or Rosenbaum’s mistake, or whatever other variant), their product will suffer, and readers will be put off their work. And if writers want to stay sharp and produce work worth reading, they’ll learn to correct for whatever is suffering in their process. In some ways, it’s a self-correcting problem.
I'm so glad you addressed this topic as it is so incendiary. One of my favorite authors has a Facebook page where if you even hint you like some aspects of AI you are immediately banned. Thanks for giving me the space to comment on your thoughtful essay. "The only metric that matters at some level is our personal appraisal of the words themselves." You have given me a way of looking at AI that makes sense and I can use as a guidepost. Thank you.
Joel, I hope you won't mind this question: what is your source for the quotations from Tokarczuk's appearance at Impact CEE 2026? I haven't been able to track down the recording from the event yet, only an article about her event (in Polish).
I linked it in the piece above. As a non-Polish speaker I’m reliant on the translation provided. I figure it’s directionally accurate or the anger about the revelation wouldn’t have been so sharp, nor her response so quick.
Thank you! That's a piece from Telewizja Polska, and it's trying to be balanced, but there's so much nuance lost when what is recapped is a conversation from a live event. I would take issue with the "beloved" "quote" -- there's a Polish convention there that conveys humor and is more like an older diner waitress calling a customer "love" than addressing someone as "Beloved." In Polish, the joke is apparent. In this translation, it's not.
Authors can do final edits of their books, yes. But an expert objective view also works well- editors specialize now in different types of edits to prepare a manuscript. The trouble is that editors also charge a fee for their services, which some writers cannot afford. AI has elbowed into their jobs as much, if not more, than those of writers, and clearly the results are too shoddy for them to really be a replacement.
It’s a question of tradeoffs. I’ve been a professional editor for twenty six years or so. Editors are essential to all sorts of literary work. But I like getting paid. I don’t hold anything against a writer who wants some of those benefits through another means. As a substantive editor, I’d put my judgement above Claude’s. But as a copyeditor and proofer, the old boy’s pretty good. I don’t think in the strictest sense there’s a real comparison between a human editing and a machine. But the machine can be helpful and works for less than peanuts.
Thanks for the article- I love your posts because they really make me think about the past, present, and future of literature and I guess for that matter any artistic pursuit. What is interesting to me is how a supposedly errorless entity (AI) can make such big errors. A lot of people use AI (I think) to answer questions and aid research about topics because they think it will give them the "right" answer quickly. However, it all goes back to "garbage in, garbage out". (which I reviewed to try and ensure accuracy through a Wikipedia article if I am to give my source of verification 😂) I always wonder who is feeding information into the AI and how do they decide what goes in? What happens when the individuals training these machines lack the ability to do their own research and verification of the information they are feeding it? (I think of the book Unmasking AI by Dr. Joy Buolamwini) It is a bit frightening to think that there are people who are controlling the learning of an entity who is touching the output of so many people. In the medical field I am constantly using AI- for research into the latest articles on new treatments (a human cannot possibly keep up with all the research and new revelations which are coming out in the medical field on a daily basis- but AI can) to use for patients as well as a transcriptionist when talking with patients. What I have found amusing/scary/weird is that sometimes AI hallucinates or just creates something outside of its input. For example when using it as a transcriptionist which takes my conversation with a patient and puts it into note form I have caught the AI creating things which were not part of the conversation. Once, after a telephone call with a patient, I reviewed the note and saw that the AI had "created" a limited physical exam for a patient I had never laid hands on. It said the lungs were clear even though my stethoscope and ears were hundreds of miles away from this patient. I reported the mistake but spent days worrying about it. I review every single note my AI transcriptionist writes and correct the mistakes. Most of the time it is really, really good- much better than I or a human transcriptionist could do. I used to review and correct mistakes from a human who transcribed my notes before they did the final entry into the chart so I can attest to this improvement. The AI is much better but I NEVER had a human transcriptionist "create" a physical exam on a patient. So where did this come from? How did this happen?
Getting to my point and sorry for the wordy meandering- when AI only can give out what is put in, what happens when new idea generation and creativity wrought from the organics of our own mind begins approaching a vanishing point? What happens when the machine is just fed its own already spit out ideas because everyone is cycling through the same (bad?) data they gathered while "researching" an idea or "generating" a book? It brings to mind an ouroboros.
The second, and to me more interesting point, is where do the hallucinations come from? Is this just a quirk that is happening because of programing or is there something bigger happening? Is creation occurring in AI?
Finally, one thing I have learned in medicine is that treatments and ideas which we originally thought were set in stone during medical school are often modified or debunked in about a 10 year cycle as new research comes out. What happens in the workings of the AI "mind" as this cycle of new data is inputted? Can it parse through and critically examine new articles and decide which is more likely to be accurate as the years of input move forward? In ten years when someone puts in the question- "what is the best treatment for condition xyz?"- what will AI do? 🤯🤯
One comment that you made that has a lot of impact but doesn't often get a lot of explanation is in regards to the speed at which things are changing. I have been using ChatGPT and/or Claude for research and final edit checks for a little over a year. In that time I have gone from having to reject quotes after double-checking to being scolded by the AI for missing a word in a quote that I got out of books in my own library. The change to be obsessive to the point of ridiculous about copyrights has gotten so bad that the only translations of the Bible it will give me a quote more than a few words long are translations that I couldn't really recommend. In either case for me the complete 180 on quotes and copyright infringement just continues to illustrate the limited usefulness that AI has for a conscientious writer.
Thanks, Joel! All I can say is: AI and writing: the slipperiest of slippery slopes.
No disagreement. We might come down on this differently, but that’s why the challenge I’m trying to highlight here is that, as readers, we almost have to be agnostic about its use. At this point, that’s what I’ve chosen to do. It’s already all over the place, whether I like it or not. But when I ask myself what I object to as a reader it’s bad writing—AI or otherwise. The Jeff Goins post I linked to above captures my feeling of revulsion and AI-inflected prose. When I start seeing those tics and tells I get irritated. But not because it’s AI. Rather, it’s because those tics and tells reflect lousy writing. More than reflect bad writing, they are bad writing. The product betrays itself.
"I personally don’t care how enthusiastic Tokarczuk or Qudan are about their LLMs, or even how they use them. I care about their artistic judgment displayed on the page. Is it any good?"
With respect, I think I disagree. This stance fails to grapple with the fact that writing has long had a strong social or communal character. In my own field, I see the extraordinary literary production of Greek writers in the 5th century B.C. and Roman writers in the Age of Augustus. One suspects that without Herodotus, no Thucydides; without Sophocles, no Euripides; without Plato, no Aristotle; with Vergil, no Ovid. So we have to ask: what is supportive of a vibrant community of writers who are continually influencing, inspiring, and challenging one another to higher and higher levels of achievement? Put differently, it can't be just about the product; it has to also be about the people behind the works. A writer who looks to her preferred AI—not to her friends and rivals who are nearby likewise sweating over their manuscripts—for creative sustenance is not giving anything back. In effect, her relationship to her cohort is parasitical. She wants to benefit from what that community generates, e.g., a kind of nursery for creativity and a reading public eager to buy books, without contributing to it. She thinks, "I have my AI. What do I need them for?"
That’s a valid point. But I don’t think it’s binary. Writers who use AI can still be—probably will want to be—in community with other writers. Humans need humans; to one degree or another, we like company. And yet plenty of writers have been reclusive types who self-isolate. I don’t think it’s either/or. It’s also possible something new will emerge that could prove good or bad—but we’ll have to wait to see what the outcome is.
I hear you. But one obvious follow-up question is this: when you do participate in our community, what do you have to offer the rest of us? What skills and insights *acquired by blood, sweat, and tears* can we learn from?
"I used AI merely to organize my thoughts."
"Merely? Okay, maybe someone else here did it the old-fashioned way and can help me with this."
Isn’t that self-correcting? If an author has little to contribute, they stop getting invitations or feel unwelcome in the group. They have to bring up their game to participate—which is true whether they use AI or not. No one likes working with people that don’t pull their weight.
At the same time, there may be new communities that form—probably will be since humans are incurably social—where writers who use AI find common cause. But even there, if someone comes in saying, “I used AI to merely organize my thoughts,” they’ll be shown the door too. No communities work without reasonable levels of contribution from the participants. Freeloaders aren’t welcome long.
So your revised thesis might be something like this: "As a reader, I care about artistic judgment displayed on the page. As a cultural critic of all things literary, I care deeply whether and how that discernment is acquired and would caution any upcoming writer that there are no shortcuts to becoming the next Joseph Conrad or Virginia Woolf."
I would endorse that. I don’t think there are shortcuts to judgment. That only comes from experience. Which does expose a real, potential threat with people using AI without the benefit of “learning the hard way.” For what we value right now, that’s the only way to learn it. It’s possible that’ll change—cultures evolve in ways current people can’t imagine.
This seems like a helpful way of thinking about it--including for academic scholarship where these debates are happening thick and fast, and where journals are currently trying to legislate if and how AI can be used (talk about closing the barn doors after the horses have bolted...). If your writing doesn't hold up to scrutiny because your claims are false, your metaphors are nonsensical, or your writing is the literary equivalent of Instagram face, then it's flawed, whether you got there with or without the help of AI, and moreover that's on *you* as the author (er, the generic 'you' obviously, not you as the author of this post).
Increasingly, I think that’s the only way to handle it. The process is very important for a writer, but the product is what the reader encounters. If the writer undermines their own process by overreliance on AI (the deskilling fear, or Rosenbaum’s mistake, or whatever other variant), their product will suffer, and readers will be put off their work. And if writers want to stay sharp and produce work worth reading, they’ll learn to correct for whatever is suffering in their process. In some ways, it’s a self-correcting problem.
I'm so glad you addressed this topic as it is so incendiary. One of my favorite authors has a Facebook page where if you even hint you like some aspects of AI you are immediately banned. Thanks for giving me the space to comment on your thoughtful essay. "The only metric that matters at some level is our personal appraisal of the words themselves." You have given me a way of looking at AI that makes sense and I can use as a guidepost. Thank you.
My pleasure. That’s the way I see it. I don’t have control over people’s process, but I can critique their product.
Joel, I hope you won't mind this question: what is your source for the quotations from Tokarczuk's appearance at Impact CEE 2026? I haven't been able to track down the recording from the event yet, only an article about her event (in Polish).
I linked it in the piece above. As a non-Polish speaker I’m reliant on the translation provided. I figure it’s directionally accurate or the anger about the revelation wouldn’t have been so sharp, nor her response so quick.
Thank you! That's a piece from Telewizja Polska, and it's trying to be balanced, but there's so much nuance lost when what is recapped is a conversation from a live event. I would take issue with the "beloved" "quote" -- there's a Polish convention there that conveys humor and is more like an older diner waitress calling a customer "love" than addressing someone as "Beloved." In Polish, the joke is apparent. In this translation, it's not.
I left the “beloved” thing out precisely because I figured she was being playful and it would be easy to misconstrue.
Authors can do final edits of their books, yes. But an expert objective view also works well- editors specialize now in different types of edits to prepare a manuscript. The trouble is that editors also charge a fee for their services, which some writers cannot afford. AI has elbowed into their jobs as much, if not more, than those of writers, and clearly the results are too shoddy for them to really be a replacement.
It’s a question of tradeoffs. I’ve been a professional editor for twenty six years or so. Editors are essential to all sorts of literary work. But I like getting paid. I don’t hold anything against a writer who wants some of those benefits through another means. As a substantive editor, I’d put my judgement above Claude’s. But as a copyeditor and proofer, the old boy’s pretty good. I don’t think in the strictest sense there’s a real comparison between a human editing and a machine. But the machine can be helpful and works for less than peanuts.
Thanks Joel. This was great.
Thanks! This whole topic has been interesting to puzzle through.
Thanks for the article- I love your posts because they really make me think about the past, present, and future of literature and I guess for that matter any artistic pursuit. What is interesting to me is how a supposedly errorless entity (AI) can make such big errors. A lot of people use AI (I think) to answer questions and aid research about topics because they think it will give them the "right" answer quickly. However, it all goes back to "garbage in, garbage out". (which I reviewed to try and ensure accuracy through a Wikipedia article if I am to give my source of verification 😂) I always wonder who is feeding information into the AI and how do they decide what goes in? What happens when the individuals training these machines lack the ability to do their own research and verification of the information they are feeding it? (I think of the book Unmasking AI by Dr. Joy Buolamwini) It is a bit frightening to think that there are people who are controlling the learning of an entity who is touching the output of so many people. In the medical field I am constantly using AI- for research into the latest articles on new treatments (a human cannot possibly keep up with all the research and new revelations which are coming out in the medical field on a daily basis- but AI can) to use for patients as well as a transcriptionist when talking with patients. What I have found amusing/scary/weird is that sometimes AI hallucinates or just creates something outside of its input. For example when using it as a transcriptionist which takes my conversation with a patient and puts it into note form I have caught the AI creating things which were not part of the conversation. Once, after a telephone call with a patient, I reviewed the note and saw that the AI had "created" a limited physical exam for a patient I had never laid hands on. It said the lungs were clear even though my stethoscope and ears were hundreds of miles away from this patient. I reported the mistake but spent days worrying about it. I review every single note my AI transcriptionist writes and correct the mistakes. Most of the time it is really, really good- much better than I or a human transcriptionist could do. I used to review and correct mistakes from a human who transcribed my notes before they did the final entry into the chart so I can attest to this improvement. The AI is much better but I NEVER had a human transcriptionist "create" a physical exam on a patient. So where did this come from? How did this happen?
Getting to my point and sorry for the wordy meandering- when AI only can give out what is put in, what happens when new idea generation and creativity wrought from the organics of our own mind begins approaching a vanishing point? What happens when the machine is just fed its own already spit out ideas because everyone is cycling through the same (bad?) data they gathered while "researching" an idea or "generating" a book? It brings to mind an ouroboros.
The second, and to me more interesting point, is where do the hallucinations come from? Is this just a quirk that is happening because of programing or is there something bigger happening? Is creation occurring in AI?
Finally, one thing I have learned in medicine is that treatments and ideas which we originally thought were set in stone during medical school are often modified or debunked in about a 10 year cycle as new research comes out. What happens in the workings of the AI "mind" as this cycle of new data is inputted? Can it parse through and critically examine new articles and decide which is more likely to be accurate as the years of input move forward? In ten years when someone puts in the question- "what is the best treatment for condition xyz?"- what will AI do? 🤯🤯
One comment that you made that has a lot of impact but doesn't often get a lot of explanation is in regards to the speed at which things are changing. I have been using ChatGPT and/or Claude for research and final edit checks for a little over a year. In that time I have gone from having to reject quotes after double-checking to being scolded by the AI for missing a word in a quote that I got out of books in my own library. The change to be obsessive to the point of ridiculous about copyrights has gotten so bad that the only translations of the Bible it will give me a quote more than a few words long are translations that I couldn't really recommend. In either case for me the complete 180 on quotes and copyright infringement just continues to illustrate the limited usefulness that AI has for a conscientious writer.