I am an outsider here because I haven't read or written much, but I have been wondering what others think about this topic.

1. What is the role of AI in writing?

1) AI should not be used at all.

Or,

2) AI should be used for these reasons (multiple choice):

A) Generating the entire story from a prompt.
B) Generating ideas of what to write.
C) Generating image or illustrations of the story.
D) Proofreading and making corrections.
E) Analyzing the story and giving feedback.
F) Adding details or writing parts of the story.
G) Fixing language and grammar mistakes.
H) Formatting the story and restructuring sentences.
I) Improving language such as adding adjectives.
J) Translating the story from your native language to English.
K) Summarize the story for a brief description.
L) Other, please specify.

2. When should a story be marked synthetic?

A) When AI has been used in any part of the process.
B) Only when AI has been used to generate the entire story.
C) It is complicated.

3. What are your thoughts on AI (generative LLMs) in general?

I don't think AI has any part in good writing.

This is coming from someone that started amateurly with help from AI to translate text, however I found that it severely hinders one's own ability to actually develop and craft a proper voice with my prose.

I've been trying to saddle back up writing and translating on my own without AI help and found it to be a long but healing process to regain ability to self-dictate words rather than letting an AI do it for you.

Seeing how much damage AI has done, I think we need to really pump brakes and reconsider its use.

I'm not a fan of AI with regards to most forms of Art. Where I think it is of benefit is in translation where you're writing in something other than your native language. And must admit it can help with grammar but otherwise I'd prefer not to read it.

One of my big issues with AI is it's lack of crediting the creators of the source material its been trained on.

I get your point from a reader's standpoint.

From writing standpoint, the value of your art belongs in prose and speech. Knowing the words you write rather than translating and relying on an algorithm to determine speech is a big disservice that kills voice and make your writing sound samey.

For those reasons I've been wanting to take my own stories down on this site, they sound very generic to me now that I've learned English better. I've had a couple new stories under a different pen name and they've been received well but the voice is much different.

Takes longer to write authentically but is worth it to be comfortable reading own words.

I,myself would prefare,no AI as I believe; should be written from within. At times it could provide 2(b) ideas to help promt them to come forth. 2(c) I'm not avery good drawer,so I do enjoy theillustration as an addition.2(j) it makes reading a lot easier,it certain does for me.2(k), as for the same reason as 2(j). I would love to comment about LLM's but I don't know what they are, I even checked Wiki,to no avail: give me a clue.

I tend to agree with the " no Al camp".
Somehow it diminishes the art of writing although I do accept it helps in terms of translation.

I take a very hardline, "absolute no place whatsoever" stance.

LLM means Large Language Model and is a form of AI trained on massive amounts of source material (e.g. stories) which can then be used to generate similar output

Thanks, Deeppan9; not very good about computer talk: thanks again,knowledge added for the day.

Importantly, this is done without permission of the people who created the stuff that they are trained on. So it's straight up theft.

AI is like fire. Cavemen probably feared fire when first discovered (rightly so!), but some thankfully saw the usefulness of harnessing its power rather than voting to always extinguishing it.

And usually done without their knowledge

I think that comparison is fair but with your own words, if AI is fire then all the data it's been trained on is kindling along with the capital that's being burned to keep it alight.

Unfortunately not all the data that's been burned to make it bright was taken with consent. And or course, everyone and every career being tossed into the fire for the sake of accelerationism is quite harmful as well.

Just my two cents

The human mind is so powerful. Human imagination is a wonderful thing.

If a human being can't come up with good ideas for a story without AI help, perhaps they should abandon their literary career, and concentrate on something else.....

The same also applies to any other form of creative pursuit, as far as I am concerned.

I'm sure fires used to burn down entire caveman communities and there were good reasons to try to ban it. But eventually we learned to overcome our fear and figure out how to use it for survival itself. Don't hate the tool; The onus is on our generation to harness AI's power and corral its use for good. If this is not your philosophy then you are probably on the wrong side of history.

AI isn't a binary thing and the discussion about it shouldn't be either. It's becoming integrated into everything whether we like it or not.

As a writer do you use spellcheck? If spellcheck starts to use AI to give you better contextual suggestions, would you refuse to use that AI? What about grammar check? If AI suggests an objectively more clear wording to something you just typed, would you refuse to use it? Could it not even serve to jog your memory or to give inspiration? Does the general public think that all AI can do is write entire passages for immediate publishing?

There is really no clear dividing line and it's all becoming seamlessly integrated into our lives. To categorically malign an entire technology is short-sighted. I think that finding the nuanced boundaries and insisting on ethical resource use should be the starting point of every discussion about AI.

While it isn't my place to tell anyone what to do or how they should enjoy things, my personal view is that for a story to be truly 'yours' it needs to be your own work. Things like grammar and spellchecking can be performed by a simple word processor without needing to resort to AI. Having been subjected to far too many AI-generated summaries around the internet, I'm growing to despise the writing template that ChatGPT seems to default to.

There are exceptions, of course. For example, if someone finds themselves unable to articulate their ideas in a way that they'd like, I wouldn't begrudge them using AI as a tool to help them express themselves - though they should be upfront in marking the results as the product of AI.

In terms of the stories posted here, I also feel that an exception can be made for using a few AI images alongside a story, provided they're ancillary to the writing and aren't used in place of actual description. For me, it's the writing itself that determines whether a story is synthetic or not.

More generally, the practices used to create LLMs seem to have involved piracy on a truly industrial scale. It seems ironic that, after all those campaigns to tell people not to download music and films, some companies are able to just basically pirate everything and then monetise the results.

I'll go along with 'Gloopy'. People have been writin' stories for hundreds of years.... Just use your 'Imagination',
though havta' admit 'A.I', pixs are pretty cool......

Yes, I can see all that. Call me a fuddy duddy if you like, but I'm afraid the people who AI are going to harm most are the lower paid creative types. This is an area that doesn't need the help of AI, pumping out strange clones of what people deserve to be paid for.

Messmaster, you are doing a fantastic job of steering the community through all the challenges - age ID as well as AI, and I will be ever grateful to you for this forum. As long as AI content is always clearly defined as synthetic as it is now, I'll be happy!

Sorry to be serious, but it feels like no-one on the bigger stage is fighting the creative corner.

Oh, and I never use spell checker. If there is an automatic one attached, I often reject it's corrections. Now I do sound like a moaning oldie.....

Learning to understand the core meaning of something and then using it to inspire something new isn't piracy. That's what all artists and writers have always done. Straight up copying just to rearrange and modify is piracy. The entire point of LLM's is supposed to be the former, and that's where copyright regulation needs to be enforced. Almost nobody acknowledges this, but we really need to understand the difference if we are to get anywhere in this debate.

I should mention my background first. I've written two novels (one self-published) and five screenplays (several received awards, none have sold).

I see AI as a tool, and like any other tool, it can be powerful when used properly and a menace when it's not. Currently I'm learning AI for use in images and videos. I haven't yet used it for any writing, however I do see legitimate uses for it. One very legitimate use for it is proofreading, but I can see a variety of other uses, such as serving as a critical reader looking for weak points or contradictions. As a fiction writer, I'm curious whether after a novel or screenplay first draft is completed, if AI could identify places where dialog is not appropriate for a character. It's an interesting idea. However the idea of using AI for the actual writing or original ideas is abhorrent.

When the typewriter was first invented, many writers didn't see it a legitimate creative tool. I expect a similar evolution in thought will happen with AI as we learn where it can be legitimately useful and where it's just a crutch for a lazy writer.

This is a valuable discussion. Thanks for starting it

That's a fair point. After all, I've taken inspiration from movies and various tropes in the stories I've written, which I guess could be seen as similar. It becomes more of a grey area (in my view, at least) depending on how the LLM processes its data set of human-generated knowledge and art, as my understanding is that it isn't truly intelligent in the way we think of the term, and instead mimics what an answer to the query should look like. That said, the technology is moving so quickly that my limited understanding of it might be out of date by now.

My thoughts are split between how AI was developed and what it cost (ethically, environmentally etc) Vs how it used an isolated tool. Overall I don't like AI and how it's used at the moment.

As simply a writer tool (for the purpose of this thread) I don't have a problem with people using it. Others have said its not a substitute for ideas and id agree with that but I don't think people use it for that. I think people have the idea and don't know how to convey it and a lot of them would probably never write anyway and this is a good starting point for them. In that sense, it can be good.

Personally I tend not to use it. I've experimented a little but I think AI writing is normally reads pretty obviously as AI (maybe that's bias and perhaps I've read a lot of AI without recognising it as such). When experimenting I ended doing a fair amount of editing after to get the scenes back to what I actually wanted while tweaking/removing excessive dialogue or adding additions/jokes/explicit WAM.

By the end, it probably was faster and easier but didn't feel much like it. There was a series I did elsewhere recently that was audience suggested OCs and involved a mechanic that became very samey to write each episode and I played around with using previous episodes as template to power through. In that case it was helpful (though I still edited) and it let me get to the part of the game/story people actually care about and what I wanted to write.

For my pure narratives, I don't use it anymore as not worth the squeeze plus my general dislike of it

Sorry MM, and I post this with as much respect as possible, but the "whether we like it or not" and the "wrong side of history" remarks I find really arrogant and dumb.

If you want to take the perspective of the history of technology, so be it. There are such things as bad technologies: the nuclear fucking weapon would be a good example. But more than that - any technology in the wrong hands, in terms of who wields the power in its implementation, can be detrimental. And right now, AI is fully in the hands of right-wing billionaires who are foisting it upon us in an effort to increase the gap between rich and poor. Actual supervillains.

I concede that you're right about spellcheck, but that's a category error - that's people just labelling everything as "AI". What people - especially creative people - are objecting to, is the mass theft of other people's work, to power engines that overwhelmingly produce soulless, inane content. (Oh, and having a significant impact on our already-fucked environment while they do it). Your entire site relies on the input of creative people - I really think you should take more time to understand their concerns, rather than saying "blah blah complicated it's the future sunglasses emoji". It's intolerable.

Anyway, onto writing. After that it might seem strange that I don't fully take a zero-tolerance approach, and I do think there is such a thing as ethical AI use. But again, everything is called AI now, so I need to be specific.

There's a rich tradition of all kinds of artists using new technology to inspire ideas, or to feed *their* ideas into unpredictable models or machines and to see what comes out. That can be a beautiful synthesis.

I also *get* but don't *love* its use as a tool in certain situations - if you've got the Blank Page Paralysis, you might ask an LLM for 10 off-the-wall ideas to start you thinking, like Eno's Oblique Strategy cards (but just look up the cards instead). If you want to work on a specific writing skill, like being more punchy, you might request an edit and see if you can learn something from that. (But just ask someone else instead). Those things to me are understandable, probably have minimal environmental impact, and though they rely on stolen work, that ship has sailed.

But if you're using any kind of generative AI to do the work of *writing* - to put one word down in front of another, to build a house brick by brick - then you're not writing, *it* is.

Insert "here we go again" meme here, I guess...


--PART 1: the actual subject--

Do you have sources to back these claims about how the adoption of fire as a technology occurred? (For that matter, have you done a lot of research about technology adoption curves in general?)


Don't hate... thumbscrews? Don't hate... napalm? Are you sure this is a good rule?


What determines which "side of history" is right?


This is a very discomforting message for a key moderator on a fetish website to send. I know you know the importance of consent, and I know you believe in consent as an ethical principle in at least some settings. I don't understand why this would be an exception.

(Also, it's not becoming integrated into "everything." If you want us to avoid binaries, please model that same behavior and try to keep your quantifiers to an appropriate level.)


Forgive me, but how much creative writing have you actually done? (Also, again, if you don't want the conversation to descend into false binaries, please try not to use "all" or "always" when the words "all" or "always" don't apply.)


Is this a legal distinction you're making? Is it a philosophical one? Is it just a social convention? If one musician covers another musician's song without changing it into something fundamentally new, is that piracy or an homage (or just a not-very-good cover)?


Is it, though? What do you even mean when you talk about "learning" or "understanding" or "inspiring"? Can you tell me how the correlational modeling systems at the core of LLM technology qualify as any one of those things?

I'm not asking these questions to be a jerk. (I guess I could've just called you dumb if I wanted to do that.) I'm asking these questions because the essence of your argument, as I understand it, seems to be something like this:

"People have always feared powerful new technologies, even when those new technologies were good all along. This is just that. As long as we learn to use it for good, whatever tradeoffs we have to make will ultimately be for the best."

But I don't think you've presented anywhere near the level of evidence that would be needed to make any one of those claims, let alone all three. In fact, at the moment I don't honestly believe that you've even done the research to possess the requisite evidence. If I'm wrong about that, I very much hope that you'll produce your evidence. But if you don't have the evidence, then you must have some other motivation for making these claims, and in that case I very much hope that you'll tell us what your actual motivation is.


--PART 2: disclosure of personal bias--
Like Woods, I'm also a writer. I've written four novels (one self-published, one published by a small press, one currently under review at a different small press, one in edits). I've had op/ed-style nonfiction articles published by a handful of magazines, plus one research article that was published in an academic newsletter. Also, David Graeber quoted me in a book one time. That was cool.

I don't use AI for any part of the process (no, not even to suggest words or phrasings as part of spell- or grammar-check). I never will, because, speaking as someone who does actually do creative writing, the process is the point. To me, using AI in my work would be like having a robot chew my food so that I can swallow meals really fast without tasting them. It's not even slightly what I want.


--PART 3: back to the subject--
In brief, my position is that the use of LLMs and other AI-like tools dangerously erodes important cognitive, emotional, and dispositional skills. These tools do allow for acceptable-quality products to be created faster and with less work, but I think there are three reasons why this is not a worthwhile tradeoff:

1. We don't need more acceptable-quality stuff. We already have more than enough of that.
2. As far as we know, it's important for individuals to have the skills that are eroded by the use of AI.
3. We have no real idea of what a society would look like if the vast majority of people lacked the skills in question, but we should probably err on the side of not going down that road.

Hopefully the skills to which I'm referring are still prevalent enough that we can engage with these ideas. I guess we'll find out.

This is such a tired argument already. God has the AI slop bubble really been going on this long enough for the arguments to be worn out? Homage is one thing, but that comes with deliberate intent and natural human use of inspiration. Straight up stealing people's work without permission and then copying it is very different. Tech bros just went with the "ask for forgiveness, not permission" angle.

Put it this way: if LLM training is not stealing, why did they bother to do it covertly/secretly in the first place and why are lawsuits popping up all over the place?

AI definitely has some uses elsewhere. A friend of mine kindly explained how useful it can be with many disabilties, but OP's question was about creative fields and I staunchly belive that it has NO place there whatsoever. " AI cannot be spiteful or horny. Therefore, it can never produce true art."

AI is for when I want a "Choose your adventure" type interactive experience. I've gotten so familiar with the typical AI slop language and style that I won't read it outside of that.

AI is way too hyped and is dissapointin seeing the amount of yall on here that jump on that tech bro bandwagon. Anybody that tries to say AI is the future frankly got no self respect, awareness and have no understanding of computers as a whole.

First, no shit AI always had functions for data and computing. That's what it's designed for and thats where it belongs. Forcing to use it in all areas of life that ain't those is all marketing and kool aid them venture capital bros push on you because the economy can only run if people be consuming.

Second, AI wishes it can touch a fraction of what our brains can produce. They need to hog up all the land and power and water meanwhile chump shit still hallucinates and provides bad data.

Third, any of yall tech bros that think AI is good always love bringin up profound terms, trying to tie up "HISTORY" or the "FUTURE" trying to get all Macro with shit, meanwhile the PRESENT and CURRENT reality of the fucking situation is fucking sour and deplorable.

AI companies havent been profitable especially here in USA since them billionaires subsiidize it with taxpayer money (By the way, social security about to be insolvent cuz of that shit, so good luck to all them boomers that love AI).
https://isaiprofitable.com/

The only people making money from AI is literally NVIDIA lmaooo. Check the sheets thats why bitches like OpenAI be scared to have an IPO, and the most recent one that did was corrupt as shit (SpaceX) which ironically again, the section that is AI (xAI) is the BIGGEST LOSS on their fukcing balance sheet.

But oh well, yall fuckaz thing we non-Ai people are on the "wrong" side of history. Let AI rot yall brains (https://www.nature.com/articles/s44387-025-00063-1) or destroy the world around you (https://www.gao.gov/products/gao-25-107172)

Let me remind you that history is relative. The Native Americans them white boys genocided was on the
"wrong side of history". That kinda talk is the kinda shit priveliged people say to make themselves feel good about the horrific shit they into. Damn right Imma be on the wrong side of history all day if it means I can keep my soul and not become them people willing to ruin everything to pat myself on the back because I think the future will improve. Fuck that, live in the present reality of life.

I use AI quite regularly for personal roleplay and expanding on ideas, but... I genuinely hate reading stories written by it. It may be the settings and instructions I've given it, and may not be the case for people who better understand how to properly set up story-writing prompts, but most of the AI stories I've read or role-played are completely ridiculous, with horrible, unrealistic dialog. ChatGPT and Gemini's base writing styles are very easy to pick up on, and I'll often stop reading other people's stories once I realize it was written by AI.

When I'm reading through its responses during story-writing or role-playing, I'll often skip over entire paragraphs of unnecessary stuff (and usually the last paragraph or two in each response where it tries to wrap up the story with some nonsense moral-of-the-story or happily-ever-after BS).

I've taken a few of the stories I've written with it and completely rewrote them in my own words, occasionally keeping an AI-written sentence here or there if it sounded ok and fit my writing voice. I've uploaded one of them on here a while back, and the original AI-assisted version was so bad that I almost quit "translating" it into my own words a few times.

I see AI as a useful tool for bouncing ideas off and helping to flesh out concepts, but can't stand it in the finished written product.

On the other hand... AI images to illustrate the story? I love them as long as they fit the story (including outfit or scenery details) and enhance the experience. Badly-done or inaccurate AI images can have the same effect on the experience as badly-written AI stories.