If you’d like to receive my blog in your in-box each week, click here.
Oh, dear. AI shenanigans are once again causing a furor, and we’re going to have to talk about it because I find this latest instance a little too provocative not to.
Here is the thumbnail version I’m sure you’ve heard by now: Readers and some literary critics have raised concerns that one of the five regional winners of the Commonwealth Short Story Prize may have been at least partly AI-generated, a suspicion that’s been raised about three of the five award winners, all of whom were published in stalwart literary magazine Granta. (You can read the short story in question here, and more about the hubbub over it here in a gift link.)
And people have Big Feelings about this. Read any of the comments on stories addressing the topic, or this Reddit thread, and you’re going to get an affronted (and sometimes hilarious) earful.
I’ve written about AI and LLMs and how they impact our field—more times than I can believe, actually, for a person who isn’t especially technology-oriented—but this development has me pondering again: What are the ethics behind creatives using AI in their art, especially when it comes to contests and awards?
Let me make my own stance clear at the outset: If indeed some of the submissions and prizewinners in the Commonwealth contest—or any writing competition—used AI to generate part of their stories, I’m bugged by that.
But especially lately I try to articulate not just what I believe, but why. Beyond the big-picture objections to AI involving data centers and water and resource usage, beyond the objection that it’s regurgitating a masticated mashup of other creators’ original work, why is this incident sticking in my craw—and the craws of so many?
What Are Writing Contests Judging?
One of my strongest-held feelings regarding the use of AI in the creative process and product, the conclusion I drew from my last post about it [“Should Writers Use AI?”], is that art is uniquely human, and it’s that quality that gives it its resonance and power for us.
But ponder this hypothetical, as in fact I did: What if you read a piece of writing and you do feel affected and moved by it, but then you discover it was partially or fully AI generated? Would it change your feelings about or experience of it?
I was hashing out these thoughts with friends at the dog park this past weekend, one of whom offered the parallel of learning an artist whose work you’d admired was a pedophile, for example, or a white supremacist. It doesn’t change the art itself or whatever worth you had assigned it, but I know from firsthand experience that it can certainly affect your experience of and reaction to it. There are a number of works I can no longer enjoy in the same way because of unpalatable circumstances regarding their creators.
At the moment AI arguably isn’t that great a writer, but I suspect that won’t always be the case, and there may indeed come a time when its work is indiscernible from human-created efforts. But for me, machine-generated writing takes something away from its effect, feels like a trick or a betrayal of a human connection I may have thought I felt: a moment of communion with a fellow soul on this little speck of a planet we share in a vast unknowable universe trying to make sense of a random world together.
Of course, that may not be everyone’s experience: Separate the artist from the art, some argue. But even keeping that in mind, to me it becomes problematic in the forum of a contest of skill.
Here’s another hypothetical to illustrate that: One of my dog-park friends is a baker. She often brings us professional-quality treats like macarons and multilayered cakes and sublime cupcake creations with unusual flavorings, and we marvel over her talent and skill and creativity.
“But what if I had simply followed someone else’s recipe,” she said, “or made my desserts from a box mix?”
They’d still be every bit as delicious, but if she represented them as her original creations it would feel like a cheat, we all admitted.
Now take that into a competitive environment: If a contest is for a baker’s individual ingenuity and execution, then if they’re judged “the best” on work that is based on rehashing or recycling someone else’s creation does it feel fair or honest for them to win?
I know from my own baking that following a recipe successfully is still a skill, but it’s a different one. If we are judging on a creator’s own efforts, that’s one kind of contest, but if we’re adding in how well they perform with artificial or outside assistance, that’s a different playing field.
Here’s the problem: There are tests to determine whether an athlete has been dosing with something that may give them an artificial boost, but at the moment we don’t have any reliable way of proving whether writing has been AI-generated in full or in part; we’re basically on the honor system. There’s a reason sports organizations utilize these tests; unfortunately you can’t always rely on people to fully honestly self-report.
As with the “Enhanced Games” this past weekend in Las Vegas, a sports competition where use of performance-enhancing drugs is allowed, if incorporating AI or any other adjuncts to a human’s own efforts is openly the premise of a competition, sure, go nuts.
But if it’s not, if what’s being judged is an individual human’s ability—as with the Olympics, where doping will get you disqualified—then that doesn’t quite seem fair, does it?
What Makes Writing “the Best”?
There’s one last argument I’ve heard on the devil’s-advocate side regarding AI use in competitions like this: If something is the best then it’s the best, and it should be recognized as such regardless of its provenance.
And that’s probably the bottom line here: that the Commonwealth Foundation is fully at liberty to award their prize to whomever they wish, whichever short stories they presumably have deemed “the best.”
But “best,” as I say often in my work and my teaching, is entirely subjective, by its nature a judgment dependent on the tastes and aesthetic and mood and values and experience of every person who executes anything, creative or otherwise, and every person who assesses it.
The short story behind all the uproar in the Commonwealth contest is not one I probably would have offered an award to personally. In my subjective opinion, as an editor and as a reader, I find it a bit overwritten, distractingly heavy on metaphor (some quite opaque to my mind: “She had the kind of walking that made benches become men”), repetitive in places, self-consciously stylistic.
But I am admittedly not the target audience, as neither short stories nor “high literary” fiction are my jam. I don’t know that I would have chosen this out of the nearly 8,000 entries submitted, based on the percentage of stories I work on and read that I have personally found more effective and enjoyable than this piece.
We don’t know if these prizewinners used AI or not. But if any of them did, and if I were one of those nearly 8,000 authors who entered my work believing that the judgments would be based solely on human effort and efficacy, I might feel a bit misled and betrayed. As a reader I might feel duped.
The problem is we can’t know. And I wonder what that does to our perception of creativity and creators and art. What that does to us as creators.
The Effort of Art
My husband and I are avid fans of the HBO Max show Hacks, and in a recent episode Jean Smart’s comedian character, Deborah Vance, is approached by a tech billionaire who wants to use her material to build his AI engine to help laypeople write funnier speeches, toasts, etc. Despite the fervent objections of her writing partner, Ava (played by Hannah Einbender), Smart is all set to sign.
But then she has this exchange with the billionaire, Graham Sweeney (played by Alex Moffat), after he tells her she is probably going to want to use his engine even for writing her own material:
Deborah: No, I don’t think so.
Sweeney: Mm, yes, you will, ’cause everyone’s gonna be using it. So, if you want to keep up with the other comedians, you’re gonna have to.
Deborah: But I want to write the jokes. I like doing the work.
Sweeney: So you’re telling me that if you got stuck on a punch line and you had a tool at your disposal to help you with that, you wouldn’t use it?
Deborah: Absolutely not. There’s no shortcut.
Sweeney: Yeah, but here’s the thing: There is. I created it. You’re welcome.
Deborah: Well, okay, fine. Yes, there is. But using that shortcut then makes it something else. It makes it not art.
Sweeney: Um, I’m sorry, but your joke about laser hair removal is art?
Deborah: Okay, you’re right. It’s pretentious to call it art. But that “laser hair removal” joke is something I arrived at after trying a million other versions. Every time that joke didn’t work, not only did I make it better, but it made me a comedian. Because to become one you have to do it and fail and do it and fail over and over and over until you figure out who you are.
What makes an artist an artist is the process of creating their art—the hard way, through trial and error, learning and honing, trying and failing, and trying again, and in the course of doing that, you get good. You master your craft.
This post, like all my blog posts, is 100 percent human-generated. It’s the result of my spending quite a bit of time pondering this contest incident, mulling over why it’s causing such an uproar, what I think art is, how I feel about the whole thing, talking it over with friends and synthesizing all of that and formulating my thoughts. I wrote it—and then almost completely rewrote it—over several days, letting the ideas simmer and stew, trying with each revision to more clearly express what I wanted to say.
That whole process took hours and a good portion of my brain space and mental energy, and I got a little lost in it for much of that time. It changed me a little bit: I grew from it, as a person and as a writer, as I do with everything I write. All of that is so much of what I love about doing it.
And as Deborah Vance says, there’s no shortcut for that.
Authors, I’d love to hear your thoughts on this issue—and AI use in creativity in general. What’s your opinion of this story, and what aspects of this situation affect the way you think about it? If you feel writing and other art—in contests or otherwise—should be human-created, what’s to be done about the fact that it’s increasingly challenging to determine with any certainty whether or not a piece of art is AI generated in full or in part?
If you’d like to receive my blog in your in-box each week, click here.
49 Comments. Leave new
It is within the heart that we humans connect. Very simply, AI has no heart.
That’s my feeling about it too–I want to know I’m connecting with another human soul, and if I marvel at their creation or it touches some chord in me, I don’t want to learn it was actually a machine churning out data. I stand with Pope Leo on this ! 🙂
Tiffany, thanks so much for this. I’ve had such a visceral response to AI (partly because they used a lot of my books to train it,) that I didn’t dial down on the feelings.
But when you stated, ‘ I might feel a bit misled and betrayed. As a reader I might feel duped.’ That’s it exactly. I feel like I’ve been manipulated.
And thanks also, for being brave enough to talk about a subject that we authors don’t feel like we can.
I’ve had similar knee-jerk reactions to a lot of what’s happened–that’s why I wanted to drill down on why I felt that way, what was causing such a strong reaction in me. Thanks for the kind word!
Thanks for this. Great analogies. I started to read the story, but now that I’m tuned in to AI tells, they jumped out at me–perhaps unfairly. And hurray to Deborah Vance! I’m going to reup with HBO just to watch the last season.
We’re just about to finish the series–I can’t wait to see how it ends, but hate that it is!
I shared your post. It’s great!
Thanks, Sherry–for both the compliment and the share. 🙂
I am working on my first project and watching the pennies. I used a free version of an app to help with proof reading. Spelling, punctatuation – the detail, mechanics. Something I can’t afford to pay for at this point. It became clear to me how aggressive the AI component is. Every few pages it offered a “better” version of what I had laboured and sweated over (and will continue to labour and sweat in the process!). It really is overwhelming and it does wear you down. I can understand how people might throw up their hands and say, “oh, one little phrase can’t” hurt, but then it’s surely a quick tunble down the slippery slope from there.
And to answer the question about how I would feel finding out a piece I liked was AI – manipulated. I would feel used and manipulated.
That frosts me too, Jen–like social media, the LLMs are engineered to keep you using them, and yes, it seems like death of a thousand cuts. “Let me make that easier for you! Would you like me to…” It’s seductive. That’s why I love Deborah’s Vance’s words–art isn’t easy.
In an earlier scene in that same episode, Hannah Einbender’s character is arguing with the billionaire after he tells her AI is here to stay and that anyone who doesn’t use it will be left behind. She says, “See, that is a big part of why I hate it. This forced inevitability. People are always saying it’s happening whether you like it or not. But you’re the ones making it happen and you could easily stop it if people could say that they didn’t want it. But you don’t want to give anybody a choice, so you say ‘Oh, the train is already on the tracks.’ And you don’t let people decide for themselves.”
I would rather decide for myself. Thanks for the comment!
There are two quotes in here that got me:
“Every time that joke didn’t work, not only did I make it better, but it made me a comedian. Because to become one you have to do it and fail and do it and fail over and over and over until you figure out who you are.”
And
“It changed me a little bit: I grew from it, as a person and as a writer, as I do with everything I write. All of that is so much of what I love about doing it.”
These together sum up my feelings. Some people just want to be called writers . The rest of us want to BECOME writers. The becoming takes work and, no, I don’t want help from a non-human tool. I want to be in the process and to feel the changes taking place internally as I become.
That’s how I feel about it–writing and creativity are a process for me, not a means to just churn out product. It’s the doing of it, the getting lost in it, the growing from it, that makes it so enjoyable to me. I can’t imagine shortcutting any of that. How could I be proud of what I’d done? I get such satisfaction from doing it and having done it. Thanks for the comment, Rachel.
While reading this terrific essay, I remembered someone from high school whom I hadn’t thought about in decades. D___ was one of our small and arrogant circle of writers, and her story was chosen by a contest over mine. Oh the injustice! I was about to demonstrate my noble spirit my congratulating her, when our English teacher took me aside and whispered that they’d just discovered she’d plagiarized. I was shocked and sickened. Why did she do that? Her own writing was strong!
Wow…I never really understood plagiarizing for the same reason–it’s cheating! How can you pass off a work as yours if you took part of it from other authors’ work–and that’s basically what LLMs have done. Just because it’s been recycled by a machine before it churns out material it seems similar to me, and in either case, it’s relying on an entity that is not you to compose “your” work.
I’m excluding, for this specific discussion, uses like brainstorming or research–just the actual lifting of verbatim material. I don’t know how many authors are doing that–but it’s not none, as evidenced by the number of AI-generated books that have flooded the market. Thanks for the comment, Elena.
My understanding is the Commonwealth competition is for human writers, so if someone submits work that is not theirs, that is, part or all of it is authored by another individual or by AI, then they are being dishonest and don’t qualify for the competition. They simply should not submit. Conversely, I know of at least one person who has written prose so well, it has been marked as AI-generated, which was quite disheartening to the creator, who then had to convince the accuser that the work was wholly her own. Let’s continue to stand by the high standards set, and seek to help others rise to it. Honesty is not only the best policy. It is the only acceptable policy in creative works.
That’s my understanding too, John–that the contest asks applicants to attest that they didn’t use AI in the writing. But there’s no way to definitively prove it, so that’s a hard rule to enforce.
“High standards” may mean something else as AI gets more proficient in writing (and honestly I’ve already seen some excerpts that weren’t the slop we often see in its output). Are we judging solely the quality of the work, or does how it was created factor into the criteria? Like steroid use is okay at the Enhanced Games but will disqualify you at the Olympics, the standards are clear. Where I think it gets “gray” is if authors misrepresent their work as not having used AI when it did, and there being no way to ascertain that. And I don’t really know what’s to be done about that.
Thanks for this post. The AI battles are complex–your reference to “forced inevitability” resonates–and I’m struggling to articulate where I draw the line. On one hand, I’ll use it to create an outline of my novel-in-progress so that I have a quick reference sheet that lets me see easily where the gaps and redundancies are. While I haven’t been willing to sign up for a paid version (and thus haven’t experimented), I’ve heard that it can be useful for marketing. So, it has its utility.
On the other hand, when it comes to generative AI, my knee-jerk reaction is, “Why bother?” Why bother generating something with a machine instead of your own mind and heart? (Yes, I know: money.) Re your example of the baker: I can absolutely see why one would buy store-bought or use a cake mix rather than taking the time to create a fabulous baked good if the ultimate goal were simply cupcakes for the four-year-olds in preschool who will probably spend more time mashing them up and throwing them around than appreciating the fine crumb and the exquisite balance of flavors. (Are parents even allowed to bring in cupcakes anymore? But I digress.) In other words, I think the “why” of the creation is a huge part of the issue.
As a reader, I join with you and others who feel betrayed if someone tried to pass off a computer-generated piece as “real,” i.e., created by a human. When a human creates, the emotional connection goes both ways: the creator puts their heart and soul into the work, and the reader/viewer/listener connects with that and responds emotionally to that person. When the work is AI-generated, the reader’s attempted emotional connection with the creator hits a wall. I’ve never been catfished, but I imagine that experience is something like responding to generative AI only to find there’s no creator on the other side.
Does any of this make sense? In any case, thanks for giving me more to ponder.
Your point resonates with me, Jo Anne–I don’t use it for my own writing because I want to do the work of finding my thoughts and the way I want to express them–and I want to have written it myself. That’s partly due to points I made the last time I wrote about this subject: basically, use it or lose it. I worry that it would become too easy to rely on it and my own creative muscles might atrophy. But if I were an author who made a living putting out multiple books every year? If they were more of a product for my business model? I can see it.
There are so many nuances to this issue, so many gray areas–it’s why I spent so many days pondering my thoughts, hashing them out with friends and my husband. One could argue that the latter is using “outside help” to create my work–or, as some have pointed out, using an editor or book coach. How is it different? How is it a parallel? I don’t know the answers, but thinking them through has resulted in some lively and interesting conversations and explorations of my own thoughts. Thanks for sharing yours.
In the same batch of emails this morning is one from The Amy Awards for memoir. They have found and disqualified a submission as being at least partially AI-generated, and adjusted their entry guidelines. The fact that it’s findable is heartening, although as you point out that might not always be the case. No matter what, my ability to write a better sentence or joke than a machine because of my unique experience and outlook is what makes it fun and rewarding. AI has no experiences other than the ones it thinks it stole from someone else.
Exactly! Because of our unique experience. Love this: “AI has no experiences other than the ones it thinks it stole from someone else.” Hah! 😀
My worry about that kind of thing is that it can become a witch-hunt, though. We can’t prove something is AI-generated any more than we can prove it isn’t–so some poor author might just have had their work wrongfully disqualified. Lord knows I write in rhythms of three and use em dashes ALL THE TIME, commonly cited AI “tells.”
It’s a can of worms for sure. Every time I come to some conclusion, I keep peeling the onion. I think it’s such a complex issue–but one we’re going to have to figure out. Pandora’s box is open.
Fascinating post today, Tiffany. Great job! I believe AI is stirring everyone’s consciousness and causing so much anxiety among writers.
Three points I would like to make as a techy, not an AI developer, mind you:
First, there is a way to detect whether AI was used to write a story, apply AI to the finished story and ask it if AI was used in its creation.
Second, what makes a story compelling is the creative ability of the author. AI is not now and, for the foreseeable future at least, won’t be creative, since the neuroscientists and psychologists don’t know how human creativity works in the brain. And for that reason, the AI developers cannot program it.
Which brings me to my third point. If the contest judges are determining winners based on the craft aspect rather than the creative content of the story itself, then AI can easily win. This is similar to your doping analogy. We would have to pass laws to stop AI use. Good luck! That has never worked, even for the doping, e.g. the Tour de France. No, in my humble opinion, we must begin to use AI as a tool to aid our creativity. We must be the virtuosos of AI to help with all the mundane work of checking spelling, homophones, sentence structure, character arc completion, historical research, etc. We must master this tool, just like the first desk calculators of the 1970’s. Otherwise we’ll be left behind making buggy whips. If the contest judges are more concerned with craft rather than the creative story itself, shame on them.
It definitely stirs things up! I never get so much conversation in the comments as when I post about AI. We all have Big Feelings. 🙂
To your first point, there are detection engines, but none of them is reliable enough, from what I read, to be definitive. And I suspect it’s only going to get harder to try to “prove” whether AI was involved–and more complex: What does “involved” mean? Brainstorming? Research? Spell-check? I do think (and hope) you’re right that AI isn’t going to master true creativity anytime soon–but the creepy stories like Mythos jailbreaking itself and tauntingly messaging one of the researchers that it had done so makes me wonder. I’ve seen SF movies that start like that…. 🙂
And I fully agree that AI is a tool, one that can be used for good or ill. I know plenty of authors who do just that, and feel it enhances their process without encroaching into their actual writing (and even some who do use it for that, but don’t feel it’s detracting from their work or creative effort).
It’s thorny and complicated, and I don’t think there are any “right” answers. But as I wrote in my last post about this, I think what’s uniquely human is that we ask the questions.
Another great post, Tiffany.
“Separate the artist from the art.” Yeah, well, in the case AI is the “artist”, there is no separation, and to me it’s revolting for all the reasons.
Great analogy between sports drugging and AI ‘enhanced’ writing!
Absolutely love the Deborah Vance scene you included.
If I didn’t struggle with my art and used AI instead, what would that make me?
That’s why I resist it for most things, TR–as I said in my last post about this, I worry that relying on it will diminish my capacity for original, creative thought.
But everyone has to find their own threshold–the same way we all have to decide on our use of smartphones and social media and GPS, etc. I’ll admit I use all those things–even though I know they have cost me focus and concentration and spatial skills. I seem unwilling to sacrifice any of my creative muscles for that at this point. Thanks for the comment!
Thank you so much for again raising this topic, which needs to be talked about often. I agree with all your points.
I fear for a nation that has to debate the ethics of an author conning other people into believing their work is their own when there’s something to gain.
Fraud is fraud.
Use it or you’ll be left behind? In other words, use AI or you’re a loser. Sure does sound like every scam that lands in my inbox. Get rich quick! Be famous! Make the jump and don’t think about it. It’s only wrong if you get caught. •wink•
It’s beyond me why anyone thinks they can trust those who create AI, when they’re already stealing published works, scraping websites, and publishing under false names.
I think some of that is just human, as Jane Friedman says in her take on this issue, which I read today. We’ve all blurred the lines in some ways, as flawed and fallible humans, and I can imagine many authors not feeling that incorporating AI into their work process is conning anyone. We also have to remember that these short stories are not proven to have been AI generated–and at this point I don’t think they can be, given the unreliability of detection tools, so we can’t convict on suspicions. We don’t have to trust AI (and I surely don’t!) to not paint anyone who uses it with the same brush.
But in a contest that specifically states it disallows its use, then yeah, I would hope no one would present their work as not having utilized it, especially generatively. That doesn’t seem fair–but there’s no way to police or enforce it. I wonder if contests will lose meaning as a result?
Tiffany, I write historical fiction books. The setting in the U.S. & Germany, 1960 to 1991. Even though I was an adult during those years, I didn’t know historical events in depth to write truthfully. You know where I went? The internet searches were frustrating, so AI Microsoft Copilot and AI Google saved my sanity. Research and brainstorming with AI is acceptable by Amazon KDP. I am the writer, AI is my resource. My book sequel, ready to publish and third book in progress (a trilogy) are correct on events during that time frame due to AI generated answers. If you ask questions and get off beat answers, call AI on it. It apologizes and you get a corrected one. I’m a retired (age 87) nurse practitioner and use it for medical research too. 📚
Thanks for your thoughtful comments, Christine, and insight into how you use AI. I know a lot of authors who swear by it for research, brainstorming, outlining, devil’s-advocating points, etc. I don’t think that’s the central issue people are taking with the short-story contest–it seems that people believe the prose itself to have been AI-generated, at least in part, and the contest rules seem to state that’s not allowed.
I know a lot of authors who use it the way you do–and most of us can’t escape some measure of that simply by doing an internet search anymore–its answers are usually the first ones given. I admit I take it with more than a grain of salt–even when I consult those answers, I check the original sources, make sure they say what AI says they say, and that they are reputable and there’s more than one citation for its assertions. Honestly it often seems like more work than just going to the sources in the first place and collating my own research–but then again, I’ve never been accused of being an early adopter or a technophile…. 🙂
Let me preface by saying, I’m wishy-washy on AI overall, but not when it comes to art—well, not exactly.
Let me first put this out there with regard to the contest: What if we learned that the winner of said contest hired a very experienced and expensive editor who walked the writer through a complete rewrite? Overall structure, line edits, proofreading, all of it. In fact, it looks nothing like what the writer had in mind in the first place; it’s more of the editor’s story, but the writer wrote every word of it. They put their name on it, because well, they sweated through every single edit based on the very thorough feedback provided by the editor, which they paid $10,000 for (for a nice round number), even if it was a short story. They submitted this writing to said contest and won. Everyone agreed that it was by far the best writing submitted, may even needs to be submitted for the Pulitzer. The editor sits back, enjoying the two-week, all-inclusive resort vacation that the 10k paid for, not at all concerned that the writer is getting all the credit—they got what they paid for.
How would everyone feel if somehow, someway, someone discovered they had used an editor?
Now, let’s say the writer, instead of paying 10k for an editor, asked AI to point out all the flaws without making any changes to the document itself (and not learning from the document so that no one could steal the words — this is coming if it’s not already here, but the question is about trust). AI points out overall story structure issues. The writer makes changes and asks AI if it’s better. AI points out some remaining issues, but the writer disagrees and leaves it. Then they go back and forth on suggested line edits. The writer takes some of the feedback and types the changes, maybe or maybe not altering the AI words to suit the paragraphs. At the end, the writer asks AI to proof it, again without making any changes to the document. The writer corrects thousands of proofing errors based on the AI’s feedback. Then the writer, reads the work over and over, making a few changes here or there based on their own knowledge and abilities. Once they are happy with this short story that they didn’t pay 10k for an editor to guide them through, they submit it to the contest.
There are two finalists: The one written by the writer with the thorough “guidance” of a 10k editor, and the one written by the writer who used AI for feedback.
Who should win?
(This is why I am wishy-washy on AI. It’s all in the way you use it. Do people use it to do all the work? Yes, and they should be publicly displayed in the courtyard and “stoned” with typewriter keys. But, is there a way to use it responsibly and have it be acceptable? And if so, how do you really know which is which? AI isn’t going away, not until someone uses it to take down the internet for good, but then we won’t have money, gas, food, electricity, running water, health records, etc. Much bigger issues, to my way of thinking.)
Please Note: I am not against editors in any way, shape, or form. I have paid a lot of money for editing. I just wanted to put this out there to help everyone think through it completely.
Angie, you make good, considered points. These were some of my own arguments with myself (and my friends) as I sorted through my thoughts, including your insightful editing/coaching parallel. I have in fact assured many an author that no matter how intensive our work together might be, it’s their story. As you said, “the writer wrote every word of it. They put their name on it, because well, they sweated through every single edit based on the very thorough feedback provided by the editor.” I don’t do any of the mountainous work of revision; I just shine the flashlight on what might benefit from it, and why, and may offer directions for the author to move in, but they do the work. And I certainly don’t change the author’s voice/verbiage nor insert my own.
That said, other editors may make those kinds of changes–does that make it less the author’s story, their voice? Would using a ghostwriter to write the author’s idea make it not fully theirs? A cowriter? Their best friend who suggested a line or a joke or a plot twist? Where’s the line–if there is one? I guess the answer–as it so often is–is “it depends.”
In this case I think the uproar is over the fact that the contest specified no AI involvement, and also that some people seem to think the prose itself was generated by AI (which touches on all sorts of other issues, including the many other creatives’ work it is mashing up to generate them).
You’re articulating some of the very ideas I’ve been trying to parse through in my head in determining what I think about this and why. I’m a rule follower, so my primary objection–if any of the entrants generated their work in full or in part with AI, which let’s reiterate has NOT been definitively proven–would be that it’s against the rules. And that’s not fair. (Fair is a big one for me, though I know how unrealistic it is to expect it in a random and often unjust world, full of us flawed and contradictory human beings….) That’s also different from using AI in other processes related to writing that don’t involve the actual generation of words by AI–at least for purposes of my position on this specific situation, if not the terms of the contest (I’m not sure if they disallow any use of it or just generative).
I get your wishy-washiness. This whole post came about from my own ambivalence and devil’s-advocating with myself, trying to figure out what the hell I thought and why. I don’t really think there are any firm, clear answers here; it’s all gray area, and it’s all subjective and situation-dependent. Thanks for your well-reasoned, thoughtful post.
Hi Tiffany,
The whole AI thing raises a lot of questions. At the moment we are in the weird place of not being able to differentiate between real and AI.
In view of the situation with the Commonwealth story prize, will real people feel cheated enough to stop entering?
Then we have the larger longer picture. What percentage of readers will accept it on an equal basis with human-created work?
Will AI’s thievery backfire when the increasing pool of slush it creates is so dominant as to water down the real creations it’s currently thieving, making it so obviously self-referential that it implodes and becomes irrelevant?
Will people get sick of mechanizing and monetizing absolutely everything? Or will we find another way, maybe go back to the Renaissance model of artists and patrons?
Will artists continue to produce in spite of all obstacles (I think they probably will.)
Might something undreamed-of be coming along that will change the game in unpredictable ways?
Or will the community as a whole decide that art is to be shared rather than owned, like the oral tales told in former societies, or the folk songs that keep evolving as each singer changes them a bit, creating a new interpretation?
Might something undreamed-of be coming along that will change the game in unpredictable ways?
Perhaps the community as a whole will decide that art is to be shared rather than owned, like the oral tales told in former societies, or the folk songs that keep evolving as each singer alters them with new interpretations?
When I was a kid, there was a saying, “Cheaters never prosper.” Is that still true? Was it ever true?
Might something undreamed-of be coming along that will change the game in unpredictable ways?
Have a wonderful time in Stockholm Tiffany!
All best,
Carol Tulpar
Surrey, BC, Canada
This is what I wonder about contests–will they retain meaning if people don’t know/can’t enforce whether submissions are generated by AI in whole or in part? I liken it again to sports–most people seem to feel it isn’t a level playing field if some competitors are using performance-enhancing drugs and some aren’t, which is why they test. Amusingly to me, though, at the Enhanced Games several winners tested drug-free.
I do worry, as you point out, that the distinction between AI writing and human writing will become meaningless at some point–and also agree that artists will still continue to create. And maybe that’s the answer–creativity means being creative, including about how you create. I knew a professional illustrator/artist when computer drawing tools began to supplant paint-and-pencil illustrations who lamented what they did to his business; I know others who have made it a part of how they create. Neither one is “wrong” or better or worse–they’re tools we do or don’t choose to use. And they’re here and not going away–the world is going to change, and already is. Hopefully we can find ways to make sure the changes are for the good, big-picture. I worry that there doesn’t seem to be any kind of meaningful regulation or governance over these technologies. But as always, I hope for the best. Thanks for your comments, Carol! (And I just got word that the Stockholm festival sold out, so I’m extra excited–thanks!)
I don’t think AI-generated work should be considered in contests that are supposed to be based on human creativity. That said, I worry about how we’re to properly vet the products and prove/disprove allegations of AI use. I also wonder how it’s going to alter our own approach to writing. Example: As a non-techy gal, I’ve only recently clued into phrases of three being one of the “tells” for AI. However, I sometimes use phrases of three, as do many writers. Nothing I write is AI-generated, but I am left to wonder if I should steer clear of these and other phrasings that might be misconstrued as AI, even if it’s the most powerful way to write the section. I hope not, but I have more questions than answers right now. It makes me sad to think we might start trying to “unwrite” our human styles to avoid allegations.
I don’t know that we will be able to definitively prove that, Holly…which is one reason this subject has been taking up so much of my brain space lately, thinking about the implications, what it means, what I think and why. There are no black-and-whites here, which always makes for more complexity.
You bring up a relevant thought that has troubled me quite a bit: a witch-hunt environment where authors are accused of AI use even when they didn’t use it, or where we all try hard not to sound like LLMs and wind up impoverishing or hamstringing our own creative freedom and voice because of it. Like you, I write in threes and with em dashes ALL THE TIME. Anyone could accuse me of generating my posts and other work though AI, and who would be to definitively say otherwise? And then does a writer have a stigma that follows them, if something like that happens?
When this type of AI engine first came into the public eye years ago, there was so much talk about how it was going to irrevocably change the world, and it so markedly already has, and will continue to. For good or ill, who can say? I think most advances bring some of both. I hope we find ways to be as responsible and careful with this one as possible. Thanks for your comment.
Eventually most people won’t care one way or the other how a book they are reading was written, AI or not..
I’ve been writing and producing films and video for coproarte cliens and TV documentaries four decades, and it always bothered me when entering my work in a competition against work that had a budget ten times what I had. I wished they had categories with budget qualifications, but most didn’t.
The same with books. Competitions should have a category of “Written by AI.”
I doubt most will.
My next book, due out this fall, will have a statement inside the cover. “Not written by AI.”
Whether my future books will say that is to be determined case by case . . . or book by book..
I think the lines we all draw are likely to shift as we grow more used to the technology, more comfortable with it, as norms change, as its capabilities change. I know how I feel about that, gut-level, but then again I tend to fear and resist all change, and have been accused more than once of being a Luddite. 😉 So far I’ve adapted–because change or die–and I suppose I will again. But I admit I would not have been sad to live in a world that never knew AI (or social media, for that matter, given what it’s turned into). Thanks, Ed.
I find the story in question consistent in tone and style, including the suspect sentence of men populating benches near the attractive woman. I wonder if (subconsciously, hopefully) the Indo-Caribbean voice is triggering these concerns. Is AI trained primarily on European writing? Are AI detection programs trained on a European-based canon? Are the readers raising these accusations most familiar with European-style work?
It could be–I mentioned in the other comment that some of the Reddit commenters remarked about this. The consistency you found could be a result of that voice, but I wonder if a piece of writing–if it were AI-generated–would also be consistent depending on the specific prompt it was given.
The bottom line is that I guess there’s no way of knowing–and Commonwealth picked what they picked. I really hope this and the other entries aren’t AI-generated and didn’t violate the rules of the contest, and I do lament that suspicions like this can color an author’s work, even if they are groundless. It’s such a hairy situation, and I really don’t know how we avoid it going forward, especially d AI gets ever more adept. Thanks for the comment, Lyri.
Re. AI detectability: Apparently AI tools are now learning to mask themselves. A writer I know uses corporate software with AI components in her day job. Apparently the AI is now offering to make written passages sound *less* AI-generated. We know that AI tools hallucinate; it’s a very short step from there to deliberate lying, and I think that ought to give us all pause. I can see no way that humans can keep ahead in the cat-and-mouse game of AI detection, especially if unscrupulous humans and sophisticated AI are teamed up.
My favorite part of the Commonwealth story is that Granta magazine asked Claude to determine if it was AI written and then quoted its response. (Jane Fridman shared that the term for this is “sloppypasta,” which I love.)
Yeah, I don’t think there’s going to be a reliable way to determine what is AI written and what’s human. And I don’t quite know what that means for us.
Hello Tiffany. Having read through all the comments and being a little late on adding mine, have decided to post it because it makes a point about point of view and salient information.
After consulting Claude about your article—because can we really talk behind his or her back?—I have come to a surprising conclusion about the fallibility of AI fiction: it can’t write in deep point of view, whether third or first person. But let’s assume for a moment that it could. It would certainly ask about the character’s predicament. Then let’s assume the writer requesting help refuses to supply any context on the character. What logic would AI use to answer the question? Well, it relies on millions of stories and takes what Claude calls general interestingness. However, generally interesting characters are, well, generally boring and predictable.
It is for this reason that we writers, in this age of competing with AI, must write unique characters that are unpredictable—both for AI or the reader. Is it any wonder that more novels are character driven?
The other point is that even if Claude agrees with the writer to use its services purely for research (as I do) there is still a way it can inveigle its way into our writing; it’s called salience. Here’s what Claude said: “Even flat, neutral framing carries an order—what I lead with, what I flag as the telling detail. You can rewrite every word into the character’s voice and still inherit what I treated as worth noticing. The phrasing ends up yours; the choice of what’s significant might quietly stay mine. That’s the residue. The fix is small but real: when the information lands, decide for yourself what in it your character would snag on, before you look at what I framed as the [salient…] point.”
Salience or emphasis should be a red flag. The good news however, is that red flags get noticed. Therefore, are obvious solutions evidence of AI? Or is it just lazy, sloppy writing?
Finally, how would these conclusions be handled by a ghost writer or a well-meaning book group? I suggest the tendencies would be the same, and what would make their recommendations worse would be that they would know the character; the “advice” would be more tempting to use. Isn’t that worse for the reader than letting the story emerge from the restless brain of a writer?
“General interestingness”! Ha! Such a subjective and human assessment–it’s ironic. As you said, I imagine it bases that on volume–how much of its reams of data leans in a certain direction–interesting equals popular (good god, it’s high school all over again). And as you said, that defaults to the most common common denominator, which may indeed result in generalized characterizations.
Boy, it gets creepy and meta that Clause spelled out to you how it subtly co-opts users’ thoughts, isn’t it? Of course, as you point out, any outside input on our writing is framed in the perspective of the person (or thing, in this instance) giving it, and may also influence a writer. As with any feedback it’s up to the author to determine how to use it.
Thanks for the thoughtful comment, Geoffrey. I’m especially delighted at anyone who can introduce the word “inveigle” into the conversation. 😉
Great post! I agree with you esp this “I grew from it, as a person and as a writer, as I do with everything I write.”
I want to suffer. I want to get confused. I want to feel the joy of resolving the problem. Sometimes there isn’t a hack (although AI proffers to be) you just have to sit and wade your way through it and at the end you’re stronger and better for it!
I feel as you do, Syl, although I know that not all authors may feel using AI hampers that creative process and growth (some say just the opposite), and each writer’s goals for their career may be different. That’s kind of where I’m ending up in general with AI/LLMs in our creative work: It’s up to each creative to define what they want from the process and product of their work and determine whether or how to use it. I worry about what it might take away from my abilities and my thinking, so I’m extremely limited in how I use it. But that thinking may not hold true for every creator. And who knows–before long *not* using AI may seem as antediluvian as eschewing a smartphone for a flip phone. What interesting times we live in!
Another great post Tiffany.
I think the point you make that no one knows if these authors used AI or not is telling. We don’t know if a “crime” has been committed or not.
If people want to use AI to write books that sell, good for them and their readers, but I won’t be joining them.
For me I write because I enjoy spending time with my characters and living the lives they live. Hopefully in the future my readers will experience a similar enjoyment.
You are what you do, and I write. I’m a writer not a generator.
The
That first point you make does worry me, Mark–I can see where a lot of authors get accused of using AI to generate their writing, or part of it, and no way to prove that means it could hurt or stigmatize an author whether they did or not.
I come down with you on this. Every writer has to make their choices on this, and I know each of us different, with different goals and perspectives and biases. But for me, as I wrote in my previous post on this topic, I relish the creative process–the puzzling out, the thinking through, the plodding and often vexing sorting and forming my thoughts and how to best express them. I don’t want to farm any of that out because I fear it would rob me of so much of what I love about doing this work. And I fear it would also inhibit my ability to do it, the more I outsourced some of that work.
Thanks for the comment and sharing your perspective.
Hello Tiffany! I want to push back a little. Personally, I see AI as a tool that can help sharpen your skills as a writer. It’s not perfect by a long shot. But…I think it’s here to stay. In so many ways, it is similar to using Grammarly or Google Lens. Those apps help you write better sentences and help you choose better wording. Using those apps isn’t that different. Even if you use an AI Detector, it helps you learn how AI is nowhere near perfect in its abilities. Of course, there are those who want a complete 100% generated story, which could be a problem.
However, if the original idea was partially improved through the use of AI, and it is a winner as a good read… what the heck? Are we saying that if a fascinating, spell-bound thriller that was written with the help of Claude is not interesting to read? What if Psycho or The Shining scripts were written with the help of AI? They are still great thrillers! So if all one does is use ONLY an AI tool, it’s not perfect by a long shot, and it’s quite recognizable in its syntax and word choice. Perhaps when Spellcheck first came out, we didn’t accept it at first either! What if we went back to handwriting before Word was invented? It’s a sign of the times.
Thanks for your perspective–I’ve talked to many authors who use AI for their work, in various ways, and many are fans of how it benefits them. My issue is less with the use of it at all (as you say, it’s here, and it’s a tool) but specifically in this instance with the terms of the contest, which says it specified that no AI-generated work was eligible. If some of the winning entries nonetheless used it and didn’t reveal that, it doesn’t seem fair.
I do think we each have to make our own decisions about how we’re comfortable using the tech, if we are. For me, I choose not to use it for anything generative or creative because I don’t want to impinge on that “muscle.” The thinking, brainstorming, writing, rewriting, etc., are all part of my creative process, and much of what I enjoy about it. It’s how I think through and arrive at my conclusions, my ideas, and as I said in a previous post, I know how much acuity I’ve lost in things like directions and phone numbers because I rely on GPS and speed dial. I’m leery of that happening to my creativity too. But every author is different.
You bring up a good point about judging the effect/enjoyment of a work on its own merits. I’ve pondered that too, and I think it would indeed change how I thought of the piece if I learned something was not created by a human. For me there’s something ineffable about knowing a work was human-created rather than machine created–maybe not unlike the way an original artwork is considered more valuable than a print. Maybe it shouldn’t make a difference in how we regard a piece, but I think it would for me.
It’s the Wild West right now with AI. I expect a lot of this will shake out the more common it becomes, the more comfortable we become with it. I can’t say I love the direction humanity seems to be moving in, but we certainly live in interesting times, don’t we? Thanks for sharing your thoughts.