This idealistic objective is highly commendable, but the fight could be futile. As you would need AI to do the work of detection. Then there will be another movement to do "organic detection" of "organic content". And the story goes on.
Think of interview candidates rejected by AI and employees fired by AI, or that case where a snack pack was identified by AI as a weapon in a student's pocket. This will lead to "organic decision making".
I too am open for business, for a modest fee I will arrange to meet a book publisher in nyc for a firm handshake to cement a declaration from them that they are publishing books not made with AI. I will then send a formal email saying they may publish a little gold star on their book, and my preeminence as a member of the literary elite should carry it through. I'm doing this for the people because I _care_.
What is the point of this? Any publishing house can just "self certify" that no AI was used. Why would it be necessary to have an outside organization, who can not validate AI use anyway and just has to rely on the publisher.
Writing a book is, in most cases, something which happens between the author and their writing medium, how could any publisher verify anything about AI use, except in the most obvious cases?
The one thing which matters here is honesty and trust and I do not see how an outside organization could help in creating that honesty and maintaining that trust.
I like the term “organic literature.” A significant amount of readers have no interest whatsoever in
generated prose, so there is definitely a viable market in human provenance.
An independent certification body is quite an old-world solution for a problem like this, but I’m not sure this is something that can be done mathematically. A web of trust may be all we have.
We need this for technical books. I was a chapter into something the other day before deciding I’d been hoodwinked into reading someone’s ChatGPT output
I wonder how this works since authors are more and more likely to use AI to spell check, fix wording, find alternate words, and all manner of other things. It might be useful to understand the “rules” for what “human” means.
Just another rent-seeker. I mostly choose books based on word of mouth recommendations or liking other things by the same author. This is very resistant to slop from AI and to the large amounts of rubbish that has always been published.
This is a big problem, though I would be slow to trust anyone purporting to address this problem. (Though, to their credit, this Books by People team is more credible than the bog-standard pair of 20yo Bay Area techbro grifters I expected.)
Reportedly, Kindle has already been flooded with "AI" generated books. And I've heard complaints from authors, of AI superficial rewritings of their own books being published by scammers. (So, not only "AI, write a YA novel, to the market, about a coming of age vampire young woman small town friends-to-lovers romance", but "AI, write a new novel in the style of Jane Smith, basically laundering previous things she's written" and "AI, copy the top-ranked fiction books in each category on Amazon, and substitute names of things, and how things are worded.")
For now, Kindle is already requiring publishers/authors to certify on which aspects of the books AI tools were used (e.g., text, illustrations, covers), something about how the tools were used (e.g., outright generation, assistive with heavy human work, etc.), and which tools were used. So that self-reporting is already being done somewhere, just not exposed to buyers yet.
That won't stop the dishonest, but at least it will help keep the honest writers honest. For example, if you, an honest writer, consider for a moment using generative AI to first-draft a scene, an awareness that you're required to disclose that generative AI use will give you pause, and maybe you decide that's not a direction you want to go with your work, nor how you want to be known.
Incidentally, I've noticed a lot of angry anti-generative-AI sentiment among creatives like writers and artists. Much more than among us techbros. Maybe the difference is that techbros are generally positioning ourselves to profit from AI, from copyright violations, selling AI products to others, and investment scams.
I don't care if AI wrote the book, if the book is good. The problem is that AI writes badly and pointlessly. It's not even a good editor, it 1) has no idea what you are talking about, and 2) if it catches some theme, it thinks the best thing to do is to repeat it over and over again and make it very, very clear. The reason you want to avoid LLM books is the same reason why you should avoid Gladwell books.
If a person who I know has taste signs off on a 100% AI book, I'll happily give it a spin. That person, to me, becomes the author as soon as they say that it's work that they would put their name on. The book has become an upside-down urinal. I'm not sure AI books are any different than cut-ups, other than somebody signed a cut-up. I've really enjoyed some cut-ups and stupid experiments, and really projected a lot onto them.
My experience in running things I've written through GPT-5 is that my angry reaction to its rave reviews, or its clumsy attempts to expand or rewrite, are stimulating in and of themselves. They often convince me to rewrite in order to throw the LLM even farther off the trail.
Maybe a lot of modern writers are looking for a certification because a lot of what they turn out is indistinguishable cliché, drawn from their experiences watching television in middle-class suburbs and reading the work of newspaper movie critics.
Lastly, everything about this site looks like it was created by AI.
I think for me I’m just going to accept that I won’t be reading any modern fiction, likely ever. It isn’t like there isn’t more than I could read in multiple lifetimes already out there that is pre, say, 2010. But the other side is that fiction has never been worse, because the commercial impetus to become a published fiction writer has never been lower (literally since before the 1600s, given functional literacy levels and the amount of fiction reading the average person does). The Steinbecks of the world aren’t writing novels in 2025.
So I see someone has taken Naomi Kanakia’s essay, “Contemporary Literary Novels Are Haunted by the Absence of Money”, to heart.
In fact, far from the contention that "Steinbecks" of the world no longer exist - they are prolific, and commercially successful, in a wide variety of Genres. Indeed, given Saul Bellow only published his last novel in the last 25 years, it seems somewhat callous to bifurcate the great from the good so chronologically.
Percival Everett immediately comes to mind - with Pulitzer Prize-winning James, a nuanced and insightful retelling of Huckleberry Finn, or 'I Am Not Sydney Poitier' which works almost as an homage to Steinbeck.
'The Nickel Boys' by Colson Whitehead I'd argue surpasses most of Steinbeck's more popular canon (Of Mice and Men, Cannery Row etc...). A magnificent novel in the very best of the american tradition.
'A Visit from the Goon Squad' by Jennifer Egan is probably tied for the best 21st Century Pulitzer winner with Jayne Anne Phillips' 'Night Watch' -both Steinbeck-esque in their charting of social mores in the face of an ever-changing culture rendered as the symbol and signifier drenched shadows of capitalism against the cave wall of society.
Looking at the Booker Prize since Paul Beatty, I'd also highlight 'Shuggie Bain' - Douglas Stuart's opus about growing up with an alcoholic mother in the working class Glasgow of the 1980s, or 'Prophet Song' - the requiem for a mother of four trying to preserve her family as a far-right totalitarian regime takes control of Ireland and suspends the Irish constitution.
In 2020 at the beginning of the pandemic, I set a timer, wrote for 10 minutes, live-streamed it, did it three times per day, for 35 days, and put everything unedited into a book.
It seems as if it may be more relevant in our AI writing times.
The world of literature is increasingly making itself inaccessible to broad audiences by turning this into a zero-sum game.
I wish OpenAI, Anthropic and Gemini would all figure out how to pay royalties to copyright holders anytime their content is used. I see absolutely no reason why they can't do this. It would really take all the steam out of these hardline anti-AI positions.
I've had this idea kicking around in my head now for a few months that this is an opportunity to update copyright / IP law generally, and use the size and scope of government to do something about both the energy costs of AI and compensation for people whose works are used. At a very rough draft and high level it goes something like this:
Update copyright to an initial 10 year limit, granted at publication without any need to register. This 10 year period also works just like copyright today, the characters, places, everything is projected. After 10 years, your entire work falls into the public domain.
Alternatively, you can register your copyright with the government within the first 3 years. This requires submitting your entire work in a machine readable specified format for integration into official training sets and models. These data sets and models will be licensed by the government for some fee to interested parties. As a creator with material submitted to this data set, you will receive some portion of those licensing feed, proportional to the quantity and amount of time your material has been in the set, with some caps set to prevent abuse. I imagine this would work something like the broadcast licensing for radios works. You will receive these licensing fees for up to 20 years from the first date of copyright.
During the first 10 years, copyright is still enforced on your work for all the same things that would normally be covered. For the 10 years after that, in additional consideration for adding your work to the data sets, you will be granted an additional weaker copyright term. The details would vary by the work, but for a novel for example, this might still protect the specific characters and creatures you created, but no longer offer protection on the "universe" you created. If we imagine Star Wars being created under this scheme, while Darth Vader, Luke Skywalker and Leia Organa might still be protected from 1987-1997, The Empire, Tatooine, and Star Destroyers might not be.
What I envision here is that these government data sets would be known good, clean, properly categorized and in the case of models, the training costs have already been paid once. Rather than everyone doing a mad dash to scrape all the world's content, or buy up their own collection of books to be scanned and processed, all of that work could already have been done and it's just a license fee away. Additionally because we're building up an archive of media, we could also license custom data sets. Maybe someone wants to make a model trained on only cartoons, or only mystery novels or what have you. The data is already there, a nominal fee can get you that data, or maybe even have something trained up, and all the people who have contributed to that data are getting something for their work, but we're also not hamstringing our data sets to being decades or more out of date because Disney talked the government into century long copyrights decades ago.
Books written by AI is yet another case of an application of AI that does nothing to solve existing problems that consumers have (too few books in this case) but instead focuses on the producer side of things.
Worse yet, increasing the quantity of books while simultaneously decreasing the quality just makes the situation worse for readers: more slop to filter out.
I love that they show an image of Kerouac's On the Eoad.. it's such a great example of a book that you know is based on real lived experience and that just comes to life as a wild and cacophonous daydream.
Knowing that some mad person wrote it is critical to its appeal
25 comments
[ 3.3 ms ] story [ 43.0 ms ] threadThink of interview candidates rejected by AI and employees fired by AI, or that case where a snack pack was identified by AI as a weapon in a student's pocket. This will lead to "organic decision making".
A recording of the entire process of it's creation is one possible answer (though how are deep fakes countered)
But maybe there is some cryptographic solution involving single direction provable timestamps..
Does anyone know of anyone working on such a thing?
Writing a book is, in most cases, something which happens between the author and their writing medium, how could any publisher verify anything about AI use, except in the most obvious cases?
The one thing which matters here is honesty and trust and I do not see how an outside organization could help in creating that honesty and maintaining that trust.
So this is the thing that Zitron and Doctorow are always talking about? Naked grifting in the AI industry?
An independent certification body is quite an old-world solution for a problem like this, but I’m not sure this is something that can be done mathematically. A web of trust may be all we have.
Reportedly, Kindle has already been flooded with "AI" generated books. And I've heard complaints from authors, of AI superficial rewritings of their own books being published by scammers. (So, not only "AI, write a YA novel, to the market, about a coming of age vampire young woman small town friends-to-lovers romance", but "AI, write a new novel in the style of Jane Smith, basically laundering previous things she's written" and "AI, copy the top-ranked fiction books in each category on Amazon, and substitute names of things, and how things are worded.")
For now, Kindle is already requiring publishers/authors to certify on which aspects of the books AI tools were used (e.g., text, illustrations, covers), something about how the tools were used (e.g., outright generation, assistive with heavy human work, etc.), and which tools were used. So that self-reporting is already being done somewhere, just not exposed to buyers yet.
That won't stop the dishonest, but at least it will help keep the honest writers honest. For example, if you, an honest writer, consider for a moment using generative AI to first-draft a scene, an awareness that you're required to disclose that generative AI use will give you pause, and maybe you decide that's not a direction you want to go with your work, nor how you want to be known.
Incidentally, I've noticed a lot of angry anti-generative-AI sentiment among creatives like writers and artists. Much more than among us techbros. Maybe the difference is that techbros are generally positioning ourselves to profit from AI, from copyright violations, selling AI products to others, and investment scams.
If a person who I know has taste signs off on a 100% AI book, I'll happily give it a spin. That person, to me, becomes the author as soon as they say that it's work that they would put their name on. The book has become an upside-down urinal. I'm not sure AI books are any different than cut-ups, other than somebody signed a cut-up. I've really enjoyed some cut-ups and stupid experiments, and really projected a lot onto them.
My experience in running things I've written through GPT-5 is that my angry reaction to its rave reviews, or its clumsy attempts to expand or rewrite, are stimulating in and of themselves. They often convince me to rewrite in order to throw the LLM even farther off the trail.
Maybe a lot of modern writers are looking for a certification because a lot of what they turn out is indistinguishable cliché, drawn from their experiences watching television in middle-class suburbs and reading the work of newspaper movie critics.
Lastly, everything about this site looks like it was created by AI.
In fact, far from the contention that "Steinbecks" of the world no longer exist - they are prolific, and commercially successful, in a wide variety of Genres. Indeed, given Saul Bellow only published his last novel in the last 25 years, it seems somewhat callous to bifurcate the great from the good so chronologically.
Percival Everett immediately comes to mind - with Pulitzer Prize-winning James, a nuanced and insightful retelling of Huckleberry Finn, or 'I Am Not Sydney Poitier' which works almost as an homage to Steinbeck.
'The Nickel Boys' by Colson Whitehead I'd argue surpasses most of Steinbeck's more popular canon (Of Mice and Men, Cannery Row etc...). A magnificent novel in the very best of the american tradition.
'A Visit from the Goon Squad' by Jennifer Egan is probably tied for the best 21st Century Pulitzer winner with Jayne Anne Phillips' 'Night Watch' -both Steinbeck-esque in their charting of social mores in the face of an ever-changing culture rendered as the symbol and signifier drenched shadows of capitalism against the cave wall of society.
Looking at the Booker Prize since Paul Beatty, I'd also highlight 'Shuggie Bain' - Douglas Stuart's opus about growing up with an alcoholic mother in the working class Glasgow of the 1980s, or 'Prophet Song' - the requiem for a mother of four trying to preserve her family as a far-right totalitarian regime takes control of Ireland and suspends the Irish constitution.
It seems as if it may be more relevant in our AI writing times.
I wish OpenAI, Anthropic and Gemini would all figure out how to pay royalties to copyright holders anytime their content is used. I see absolutely no reason why they can't do this. It would really take all the steam out of these hardline anti-AI positions.
Update copyright to an initial 10 year limit, granted at publication without any need to register. This 10 year period also works just like copyright today, the characters, places, everything is projected. After 10 years, your entire work falls into the public domain.
Alternatively, you can register your copyright with the government within the first 3 years. This requires submitting your entire work in a machine readable specified format for integration into official training sets and models. These data sets and models will be licensed by the government for some fee to interested parties. As a creator with material submitted to this data set, you will receive some portion of those licensing feed, proportional to the quantity and amount of time your material has been in the set, with some caps set to prevent abuse. I imagine this would work something like the broadcast licensing for radios works. You will receive these licensing fees for up to 20 years from the first date of copyright.
During the first 10 years, copyright is still enforced on your work for all the same things that would normally be covered. For the 10 years after that, in additional consideration for adding your work to the data sets, you will be granted an additional weaker copyright term. The details would vary by the work, but for a novel for example, this might still protect the specific characters and creatures you created, but no longer offer protection on the "universe" you created. If we imagine Star Wars being created under this scheme, while Darth Vader, Luke Skywalker and Leia Organa might still be protected from 1987-1997, The Empire, Tatooine, and Star Destroyers might not be.
What I envision here is that these government data sets would be known good, clean, properly categorized and in the case of models, the training costs have already been paid once. Rather than everyone doing a mad dash to scrape all the world's content, or buy up their own collection of books to be scanned and processed, all of that work could already have been done and it's just a license fee away. Additionally because we're building up an archive of media, we could also license custom data sets. Maybe someone wants to make a model trained on only cartoons, or only mystery novels or what have you. The data is already there, a nominal fee can get you that data, or maybe even have something trained up, and all the people who have contributed to that data are getting something for their work, but we're also not hamstringing our data sets to being decades or more out of date because Disney talked the government into century long copyrights decades ago.
Worse yet, increasing the quantity of books while simultaneously decreasing the quality just makes the situation worse for readers: more slop to filter out.
Knowing that some mad person wrote it is critical to its appeal