Thank you for this post, Erin. I am one of those querying novelists who spent six years on my debut. After getting some nibbles but no bites, I stopped querying, addressed an r&r and now it’s with my developmental editor for the fourth time. I’ve invested a lot of time and money, and that’s okay. As you said, writing is hard. I’m learning. It’s a business expense. But all of these authors getting cancelled, their careers ended before they’ve even started, on “suspicion” makes me feel like self-publishing is a safer route. I trust myself to usher my baby into the world and make sure she thrives.
That’s one of my worries for traditional publishing. Legitimate writers will lean into self publishing out of self care and fear. I hope you still try for an agent and a traditional publishing deal first.
Just know Mr. Falade's agents say they "cannot substantiate how the manuscript fully evolved from origin to completion." If you have drafts, workshopped your writing, and hired editors who vouch that the blisters on your fingers are keyboard related, I don't think you'll have a problem. I'm cheering you on!
Thanks for your support, Cecilia. I’ll take it! 😊 I’m not worried about proof. I have a handwritten timesheet and a spreadsheet, plus binders of drafts. Plenty proof. I’m just not sure I want to partner with a corporation that doesn’t stand by its authors. I read about authors not getting marketing or editorial support, and now we have to worry that the people who said, “Yes! I believe in your story enough to back it,” can flip on unsubstantiated (not all are like Falade) AI evidence and end your career? I think that’s unethical.
Strong and excellent stance in a sad and terrifying new world of problems. I do not understand why a writer would let anyone or anything else do their thinking. I call it the Holy Process because that’s how it feels sometimes when my brain strains or tinkers with a storyline problem and then rewards me with a solution. It’s that creativity that IS THE REWARD of writing.
I really appreciate this take. Publishers across the board need to have a clear, publicly stated stance. You see this in journalism, too. Nieman Lab just wrote about how eight Pulitzer awardees disclosed AI usage.
Absolutely, 100%, this this THIS. Without a standard, we must insist on our own. Thank you for putting my thoughts on AI so eloquently into this article.
Amen. For me, as a querying debut author, it goes beyond the pages. I worry that queries and synopses written with ai will shine over those written without (aka mine). Erin, I am curious your stance on this. I suspect it’s align with your ai views, but it’s frustrating as a debut author who’s chosen to resist the temptation of cutting corners.
I've been hosting query workshops lately and the AI written queries and synopsis stand out like a cardinal in the snow. Most newbies writing a synopsis or query either pack far too much information in or leave out too much and depend on tropes and themes. You CAN tell the difference, believe me. Writing queries and synopsis are hard work and it shows when someone puts the effort into it.
Thank you for being so clear-eyed about this! To me, using AI is the ultimate plagiarism—instead of your writing being everything you’ve read and experienced, thought and felt filtered through your own perspective to create a work of art that is uniquely you, AI takes a prompt to create a work that is defacto not you—instead amalgamating other writers’ works to create a facsimile of writing. It’s one thing to do an erasure poem or a cento or a collage poem, giving credit to the source material—that adds layers and richness to the finished poem. But to hide behind AI and say it’s you when you’ve used other people’s creative work to make a derivative commercial product is disingenuous to say the least. I’m doing an AI experiment on my substack where I’ll share two versions—I was surprised by one of the questions ChatGPT asked me during the process! Horrifying!
I agree wholeheartedly that writers should be creating their own prose. However, I have an interesting story to share on this topic.
I am a now a hopeful debut fiction writer who has worked in tech marketing. I have edited and improved commercial product stories for a living for decades. My co-workers affectionately call me “the grammar police”.
When the huge Pangram story broke this week, I put the first 5 chapters of my debut novel through Pangram, for curiosity’s sake. This was the official “version 14” of my book, and a version I hoped was ready for pitching.
Pangram came back with 85% AI for the entire 5 chapters, and a range of 35% to 95% when I later fed it each of the 5 individual chapters. Pangram also pointed to areas where it was highly confident that AI writing had been “humanized”, via attempts to alter the original AI prose.
When I fed the same chapters to Chat GPT to simply ask it “what is it about my writing that might have caused Pangram to draw these conclusions?”, GPT listed reasons including “no typos or grammatical errors” and “sophisticated sentence structure and phrasing” and “writing and pacing that is similar to successful authors in the women’s fiction genre”.
I am beyond frustrated and dismayed, particularly knowing what I do about the accuracy and dependability of early stage tech products - after 40 years in the business.
Understandable! I’m sorry this happened. I think it’s scary. That writers are going to have to defend their original pros because AI detectors are not strong enough yet. The fact that we even need to have them is the worst part. But you’re not alone I’m seeing more and more these stories.
Erin we must remember AI cannot come up with original writing. Instead it data scrapes and reorders. Prompted for fiction, AI produces already-published scenes and plots, peppered with verbatim phrases from pop songs, commercials, novels, you name it. Pangram is very good at spotting this.
I appreciate this. I think agents need to make their stances crystal clear and I do think some of this questioning needs to be up front- was ai used for brainstorming, was it used for editing, was it used for x y and z.
In the past I liked to use either ProWriter or Grammarly to help with my editing. I am a lousy speller, prone to dyslexia in word order, or in simply leaving little articles out because when I read something, the article is there. They were great tools for me. But in the past year they have jumped on the AI band wagon to offer more "generative services." It has become harder to use their services as their AI is jumping in with its "help". Why do companies like these, that were good writing "tools," think they have to jump into the generative AI universe? Using these tools will make it harder and harder to avoid the "AI" label.
I guess when subscription renewal time comes around I'll be saving money. Now to find a new reactive vs generative tool for these edits. . .
Thanks so much for this post, Erin. We toiling writers appreciate your integrity!! It's been a long road with my current project -- many years and too many revisions to count. I'm hoping this is now an advantage as I look for an agent. So happy that you and other agents see the value in all of the labor that goes into writing books!
But I'll admit that I'm afraid of someone running my hard work through AI before they even agree to read it. To have my work eaten and stolen by AI before I've even published it would be the worst imaginable outcome.
Erin: I enjoyed and appreciated your article. Where do you stand on these matters in relation to nonfiction, especially but not exclusively biography and memoir?
I think it applies for the writing. For research AI can be a terrific tool, but when it comes down to composing the text, that should be on the author, not AI, in my opinion.
I appreciate this so much and will share with creative writing faculty. As a querying writer, I've noticed that UK agencies have more strident & clearly stated guidelines. They mention their no AI policy for process work like plotting, as well as pitch letters, synopses, & all editing/revisions. I look for this clarity as that's the kind of rep that I'm seeking, as well as the reciprocal policy that agents don't use LLMs on their end with your manuscript (which can place your unpublished text in the maw of AI's training corpus forever and without consent). We need this transparent conversation in both publishing & academia.
Thank you for this post, Erin. I am one of those querying novelists who spent six years on my debut. After getting some nibbles but no bites, I stopped querying, addressed an r&r and now it’s with my developmental editor for the fourth time. I’ve invested a lot of time and money, and that’s okay. As you said, writing is hard. I’m learning. It’s a business expense. But all of these authors getting cancelled, their careers ended before they’ve even started, on “suspicion” makes me feel like self-publishing is a safer route. I trust myself to usher my baby into the world and make sure she thrives.
That’s one of my worries for traditional publishing. Legitimate writers will lean into self publishing out of self care and fear. I hope you still try for an agent and a traditional publishing deal first.
Just know Mr. Falade's agents say they "cannot substantiate how the manuscript fully evolved from origin to completion." If you have drafts, workshopped your writing, and hired editors who vouch that the blisters on your fingers are keyboard related, I don't think you'll have a problem. I'm cheering you on!
Thanks for your support, Cecilia. I’ll take it! 😊 I’m not worried about proof. I have a handwritten timesheet and a spreadsheet, plus binders of drafts. Plenty proof. I’m just not sure I want to partner with a corporation that doesn’t stand by its authors. I read about authors not getting marketing or editorial support, and now we have to worry that the people who said, “Yes! I believe in your story enough to back it,” can flip on unsubstantiated (not all are like Falade) AI evidence and end your career? I think that’s unethical.
Strong and excellent stance in a sad and terrifying new world of problems. I do not understand why a writer would let anyone or anything else do their thinking. I call it the Holy Process because that’s how it feels sometimes when my brain strains or tinkers with a storyline problem and then rewards me with a solution. It’s that creativity that IS THE REWARD of writing.
Agreed! 👏👏
I really appreciate this take. Publishers across the board need to have a clear, publicly stated stance. You see this in journalism, too. Nieman Lab just wrote about how eight Pulitzer awardees disclosed AI usage.
Absolutely, 100%, this this THIS. Without a standard, we must insist on our own. Thank you for putting my thoughts on AI so eloquently into this article.
Thank you for writing this. This makes me wonder how we can request/demand an AI policy from all publishers and publishing related organizations…
Amen. For me, as a querying debut author, it goes beyond the pages. I worry that queries and synopses written with ai will shine over those written without (aka mine). Erin, I am curious your stance on this. I suspect it’s align with your ai views, but it’s frustrating as a debut author who’s chosen to resist the temptation of cutting corners.
I've been hosting query workshops lately and the AI written queries and synopsis stand out like a cardinal in the snow. Most newbies writing a synopsis or query either pack far too much information in or leave out too much and depend on tropes and themes. You CAN tell the difference, believe me. Writing queries and synopsis are hard work and it shows when someone puts the effort into it.
Ok, that is very reassuring. Thank you!
Thank you for being so clear-eyed about this! To me, using AI is the ultimate plagiarism—instead of your writing being everything you’ve read and experienced, thought and felt filtered through your own perspective to create a work of art that is uniquely you, AI takes a prompt to create a work that is defacto not you—instead amalgamating other writers’ works to create a facsimile of writing. It’s one thing to do an erasure poem or a cento or a collage poem, giving credit to the source material—that adds layers and richness to the finished poem. But to hide behind AI and say it’s you when you’ve used other people’s creative work to make a derivative commercial product is disingenuous to say the least. I’m doing an AI experiment on my substack where I’ll share two versions—I was surprised by one of the questions ChatGPT asked me during the process! Horrifying!
I want to read this!!!
I’m working on it now—will let y’all know when it’s up!
I agree wholeheartedly that writers should be creating their own prose. However, I have an interesting story to share on this topic.
I am a now a hopeful debut fiction writer who has worked in tech marketing. I have edited and improved commercial product stories for a living for decades. My co-workers affectionately call me “the grammar police”.
When the huge Pangram story broke this week, I put the first 5 chapters of my debut novel through Pangram, for curiosity’s sake. This was the official “version 14” of my book, and a version I hoped was ready for pitching.
Pangram came back with 85% AI for the entire 5 chapters, and a range of 35% to 95% when I later fed it each of the 5 individual chapters. Pangram also pointed to areas where it was highly confident that AI writing had been “humanized”, via attempts to alter the original AI prose.
When I fed the same chapters to Chat GPT to simply ask it “what is it about my writing that might have caused Pangram to draw these conclusions?”, GPT listed reasons including “no typos or grammatical errors” and “sophisticated sentence structure and phrasing” and “writing and pacing that is similar to successful authors in the women’s fiction genre”.
I am beyond frustrated and dismayed, particularly knowing what I do about the accuracy and dependability of early stage tech products - after 40 years in the business.
Understandable! I’m sorry this happened. I think it’s scary. That writers are going to have to defend their original pros because AI detectors are not strong enough yet. The fact that we even need to have them is the worst part. But you’re not alone I’m seeing more and more these stories.
Erin we must remember AI cannot come up with original writing. Instead it data scrapes and reorders. Prompted for fiction, AI produces already-published scenes and plots, peppered with verbatim phrases from pop songs, commercials, novels, you name it. Pangram is very good at spotting this.
I appreciate this. I think agents need to make their stances crystal clear and I do think some of this questioning needs to be up front- was ai used for brainstorming, was it used for editing, was it used for x y and z.
Thank you for being the voice of reason and saying what needs to be said.
In the past I liked to use either ProWriter or Grammarly to help with my editing. I am a lousy speller, prone to dyslexia in word order, or in simply leaving little articles out because when I read something, the article is there. They were great tools for me. But in the past year they have jumped on the AI band wagon to offer more "generative services." It has become harder to use their services as their AI is jumping in with its "help". Why do companies like these, that were good writing "tools," think they have to jump into the generative AI universe? Using these tools will make it harder and harder to avoid the "AI" label.
I guess when subscription renewal time comes around I'll be saving money. Now to find a new reactive vs generative tool for these edits. . .
I am tempted to use AI to write a response. JK. You tell 'em. The scary part is the rush to cancel without a conversation.
Exactly.
Thanks so much for this post, Erin. We toiling writers appreciate your integrity!! It's been a long road with my current project -- many years and too many revisions to count. I'm hoping this is now an advantage as I look for an agent. So happy that you and other agents see the value in all of the labor that goes into writing books!
But I'll admit that I'm afraid of someone running my hard work through AI before they even agree to read it. To have my work eaten and stolen by AI before I've even published it would be the worst imaginable outcome.
To be blunt, if fiction had not gotten consistently bad, the test for AI writing would just be to read the book
Readers are partly to blame. Too. What bit of them remains.
Erin: I enjoyed and appreciated your article. Where do you stand on these matters in relation to nonfiction, especially but not exclusively biography and memoir?
I think it applies for the writing. For research AI can be a terrific tool, but when it comes down to composing the text, that should be on the author, not AI, in my opinion.
I appreciate this so much and will share with creative writing faculty. As a querying writer, I've noticed that UK agencies have more strident & clearly stated guidelines. They mention their no AI policy for process work like plotting, as well as pitch letters, synopses, & all editing/revisions. I look for this clarity as that's the kind of rep that I'm seeking, as well as the reciprocal policy that agents don't use LLMs on their end with your manuscript (which can place your unpublished text in the maw of AI's training corpus forever and without consent). We need this transparent conversation in both publishing & academia.