Crazy Taxi: World Tour announced and it's using generative AI

Quoting: Slaxer

Quoting: scaineAh, the "it's just a tool" argument. I'm not a fan of this.

Me neither, I hate it. It's the start of something that has largely only existed in science-fiction, and something that has never existed before in human history - but it's just a tool, they say. Yes, I'm aware that in its current state it's technically still not anything close to what the term "artificial intelligence" implies, but still. I feel like I'm being gaslit whenever someone says it.

Quoting: spymastermattIn particular your point about how you would make a bad table because you lack the skills, demonstrates for me exactly what I'm talking about. Here we have an actual game studio, who's professionals are using it as a tool and we are slating it because we assume suddenly that they've fired all the professionals and had people with no creative skill use the tool.

As an artist by trade, I'm not worried about AI replacing artists. That's not the problem. I just believe that the painstaking monotony of sculpting every scale on a dragon, or modeling the treads of a tire by hand, or having the discipline involved in painting something like the ceiling of the Sistine Chapel over years is what makes art beautiful. There is always an element of pain in any form of creation, and it's the secret ingredient to good art. Hypothetically, if you knew that every piece of Renaissance art in Italy was actually just made by a machine, do you think it'd be worth flying over to Italy to see? They'd completely lose their significance. AI isn't going to enable artists to be more creative by liberating them from hard work. At the very least, its use by artists is eventually going to make artists lazier and less skilled at their craft. At most, it's going to make art pointless.

I can't do art to save my life. But I do a bit of writing, and I will say that if someone tells a large language model "Write me a book about this" and thinks it's theirs because they had an idea (which apparently does happen) . . . that's incredibly stupid. Most of the ideas in a book are created in the process of writing it; you might start with a basic framework, but it's while you're knocking out paragraph after paragraph that you make it rich, developing the ideas you had, coming up with new ones in the flow of making. You'll think "OK, gonna write a few pages about this" and a couple of pages in you realize there's this other stuff that can connect in. You realize who a character is as you write dialogue and it comes out a particular way. I wouldn't be surprised if art can be kind of like that too.

None of that is going to happen if you're just getting AI to fill up the word count.

Quoting: Slaxer

Quoting: scaineAh, the "it's just a tool" argument. I'm not a fan of this.

Me neither, I hate it. It's the start of something that has largely only existed in science-fiction, and something that has never existed before in human history - but it's just a tool, they say. Yes, I'm aware that in its current state it's technically still not anything close to what the term "artificial intelligence" implies, but still. I feel like I'm being gaslit whenever someone says it.

Quoting: spymastermattIn particular your point about how you would make a bad table because you lack the skills, demonstrates for me exactly what I'm talking about. Here we have an actual game studio, who's professionals are using it as a tool and we are slating it because we assume suddenly that they've fired all the professionals and had people with no creative skill use the tool.

As an artist by trade, I'm not worried about AI replacing artists. That's not the problem. I just believe that the painstaking monotony of sculpting every scale on a dragon, or modeling the treads of a tire by hand, or having the discipline involved in painting something like the ceiling of the Sistine Chapel over years is what makes art beautiful. There is always an element of pain in any form of creation, and it's the secret ingredient to good art. Hypothetically, if you knew that every piece of Renaissance art in Italy was actually just made by a machine, do you think it'd be worth flying over to Italy to see? They'd completely lose their significance. AI isn't going to enable artists to be more creative by liberating them from hard work. At the very least, its use by artists is eventually going to make artists lazier and less skilled at their craft. At most, it's going to make art pointless.

So to draw the analogy again, is the bespoke table made by a carpentry workshop less valuable because they used a CNC to cut the top and bore the fixing holes? Did they put less creative design into it because they used the machine?

Yes there is awe in the ability to do something prior to the machines - the pyramids are impressive not because we couldn't possibly make them today, but because they were made without modern machinery - but that doesn't mean I would consider building a large scale building without the use of cranes today anything other than insane.

For me what makes good art is either the above, or mostly the choice of subject and detail.
For me the Mona Lisa is not a good painting because it took a long time and was difficult to paint, but because the choice of expression is so enigmatic and not really seen in other paintings.

I saw a fantastic sculpture in Birmingham UK made out of knives from an amnesty. It must have taken hours to put together, but that's not what made it impressive. It was good art (to me) because of what it represented.

Quoting: Purple Library Guyif someone tells a large language model "Write me a book about this" and thinks it's theirs because they had an idea (which apparently does happen) . . . that's incredibly stupid.

Absolutely agree, 100%!

But if someone develops an entire story, with a good plot and character development and twists and turns and they get the AI to actually put it into words because they just can't, and then they enrich it by reading it back and tweaking and adding and refining to make it the story they had in their head but could never express because of their ADHD, is their story less good because they couldn't write each individual word?

This, as you might have guessed, is a real story about a friend of mine who has fantastic ideas for stories, and describes great scenes from this particular book they would love to write, but can't. AI might give them the chance to do so. And yes you're absolutely right, the story has changed and refined and got better each time they've come to me over the years and said "I've changed this bit and added this bit between those scenes and tweaked this characters backstory a bit". But they're still doing that despite not writing a complete book, just snippets.

I guess my analogy in this case would be is the songwriter who writes clever lyrics, less of a songwriter because they can't sing it themselves? Should we scorn them for using a singer instead of just singing it themselves like a real artist

Again, we've taken the idea that AI can make a book given a one sentence prompt and blanket swept that that must therefore be the only way people use it

I get the idea when the tires have "flames", but what's with the green glow?

Quoting: spymastermatt

Quoting: Purple Library Guyif someone tells a large language model "Write me a book about this" and thinks it's theirs because they had an idea (which apparently does happen) . . . that's incredibly stupid.

Absolutely agree, 100%!

But if someone develops an entire story, with a good plot and character development and twists and turns and they get the AI to actually put it into words because they just can't, and then they enrich it by reading it back and tweaking and adding and refining to make it the story they had in their head but could never express because of their ADHD, is their story less good because they couldn't write each individual word?

This, as you might have guessed, is a real story about a friend of mine who has fantastic ideas for stories, and describes great scenes from this particular book they would love to write, but can't. AI might give them the chance to do so. And yes you're absolutely right, the story has changed and refined and got better each time they've come to me over the years and said "I've changed this bit and added this bit between those scenes and tweaked this characters backstory a bit". But they're still doing that despite not writing a complete book, just snippets.

Ignoring any and all other concerns: As long as they don't then imagine that they've actually written a story, and try to claim copyright to the result, I guess it would be fine? I'm sure it feels rewarding for them to get that story out. They haven't become an author, more like an amateur editor, but still, good for them.

I think the situation would be different if they'd actually written the whole story to the best of their ability, and used these tools to give them suggestions on how to make it better. A lot more effort, but there's a reason even established authors work for months, sometimes years, to write a novel, even with getting help and feedback from professional editors and readers. Most of us haven't got the talent or the patience for that, but that doesn't make it okay to just outsource the bulk of the effort.

Quoting: spymastermattI guess my analogy in this case would be is the songwriter who writes clever lyrics, less of a songwriter because they can't sing it themselves? Should we scorn them for using a singer instead of just singing it themselves like a real artist

The analogy doesn't work. Why wouldn't they be called a songwriter after writing a song? They'll get their share of the royalties. They just can't be called a singer. Just like your friend isn't an author after prompting and editing a book.

Quoting: spymastermattAgain, we've taken the idea that AI can make a book given a one sentence prompt and blanket swept that that must therefore be the only way people use it

No, what you've done is ascribe a lack of understanding to people with valid criticisms of the technology and how it is being used in the real world. Sure, some people react with their gut, but critics do include a lot of people who are actually in the know, but crucially not invested.

I know this is not the first time we've sacrificed something important, like the environment, for profit and/or convenience. But it sucks every time.

Quoting: SlaxerAs an artist by trade, I'm not worried about AI replacing artists. That's not the problem.

Yeah, the AI is incapable of replacing artists, but it's capable of chewing up all the artists' works and spitting out something more or less acceptable by an executive. So yeah, it won't replace artists but it will be used (by CEOs) as a replacement for a junior illustrator, who with some work, persistence and money (to support day-to-day living) might have grown to be an artist, something that will now be a bit harder...

Quoting: tuubiIgnoring any and all other concerns: As long as they don't then imagine that they've actually written a story, and try to claim copyright to the result, I guess it would be fine? I'm sure it feels rewarding for them to get that story out. They haven't become an author, more like an amateur editor, but still, good for them.

I guess this is where we differ. For me an author is the person that came up with the idea for the story, plot, characters, arcs, scenes, and development. So if the person (A) is doing all of that and describing them in the best detail they can to someone (B) who writes the actual words (be they computer or human) then in my mind A is still the author.

Quoting: tuubiI think the situation would be different if they'd actually written the whole story to the best of their ability, and used these tools to give them suggestions on how to make it better. A lot more effort, but there's a reason even established authors work for months, sometimes years, to write a novel, even with getting help and feedback from professional editors and readers. Most of us haven't got the talent or the patience for that, but that doesn't make it okay to just outsource the bulk of the effort.

That's a fair point and I guess then it comes down to how much of each. My friend (who is not actually considering using AI at this moment I should point out) has written several scenes of the book himself. If he uses an AI to help fill in the gaps between the scenes, is there a moment for you when it becomes not his work, or is it simply if he uses the AI at all? Genuinely interested in your opinion here, I appreciate from your later comment that my tone has not reflected that.

Quoting: spymastermattI guess my analogy in this case would be is the songwriter who writes clever lyrics, less of a songwriter because they can't sing it themselves? Should we scorn them for using a singer instead of just singing it themselves like a real artist

Quoting: tuubiThe analogy doesn't work. Why wouldn't they be called a songwriter after writing a song? They'll get their share of the royalties. They just can't be called a singer. Just like your friend isn't an author after prompting and editing a book.

I guess then the problem for me is that we don't yet have a view or word to separate the ideasmith from the author in the way we do the songwriter and singer.

Quoting: spymastermattAgain, we've taken the idea that AI can make a book given a one sentence prompt and blanket swept that that must therefore be the only way people use it

Quoting: tuubiNo, what you've done is ascribe a lack of understanding to people with valid criticisms of the technology and how it is being used in the real world. Sure, some people react with their gut, but critics do include a lot of people who are actually in the know, but crucially not invested.

I apologise that this is how I have come across, I was intending to question what feels like gut reactions but I appreciate that has come across as assuming people don't understand.
I was trying to separate PurpleLibraryGuy's suggestion of

Quoting: Purple Library Guyif someone tells a large language model "Write me a book about this" and thinks it's theirs because they had an idea (which apparently does happen) . . . that's incredibly stupid.

from the idea of people using AI to fill in gaps rather than write the whole story. From my perspective there's a big range between Wrote the book entirely myself -> Had the AI spell and grammer check -> Had the AI fix up some scenes to make them read better -> ... all the way through to "Told the AI to write a sci-fi book about an alien who lives on earth" (which I 100% agree removes any creative direction from the human)

And yes, a lot of the reactions to every announcement of AI use in any form feel to me like
"100% AI = bad, therefore 1% AI = bad"

Quoting: spymastermattSaying that a studio shouldn't use GenAI at all feels a bit to me like saying I won't buy furniture if the carpenter used a power saw

A power saw does not destroy the planet in the same capacity in a very critical time span (right now we have to solve the climate issue, every year later it may becomes the year too late to stop the dipping point chain - aka point of no return) and forces people to pay 5 times higher power bills and stealing all goods of the world without paying anything in return. There are even leaked chats of NVidia where they agreed using millions of pirated books.

All of you defenders are never ever speaking about the negative consequences beyond "a tool can be used for good and bad". And even this argument is very weak. People can also do bad things with a power saw and kill people with it like in many video games. But in reality it does more good than harm and I already know more people who died due neuronal networks (further as NN) (autonomous cars killing people) than chainsaws. LLMs are destroying much more right now than they produce benefits.

I'm tired of all these excuses and I am not even a NN hater. I would use NNs in my own game, if there is a real benefit. Prototyping or programming or generating textures/models would not be part of it. Prototyping is actually one of the most fun and creative works and bring a lot of new ideas and inspirations, I would not want to lose that process in favor of more time. That way I may would have a feature more, but all features would probably more boring. On the other hand "guessing" physic could become a huge advantage over calculating physics for using hardware more efficient and creating better visuals. These models are also much more ethical and the opposite of slop. I just wonder why all the corps oversee these possibilities.

In such a case I would explain in detail where and why I would use is. Not just to get trust back, but especially to show "it is the right way to use this as actual tool, not as sloppyfier". It is also about transparency and respect of my games community. Sadly I have not to knowledge and capacity to program these tools this way, I am more designer than programmer. And no, I don't fear NN can replace me, because it cannot do things on the cutting edge.

Edit:

Quoting: spymastermattFrom my perspective it also enables the opposite, people with good ideas for games, but without the skills / time / money to make art for it can now make their game.

People who lack skills to make a game also lack skills to do it with LLMs. I lacked skills myself a decade ago, but the engine developers deliver tools that everyone can learn to make games these days. You lack skills creating textures? Just use one of the thousands of CC0 licensed assets, you don't even need to pay money or generate them. There are experts telling you for free how to create shaders in a professional way which enables you to make the free textures look like something completely different. You still need a lot of time, even with LLMs, if it should not just become the next slop everyone hates, even if it would be handmade. And money? If you lack money you can forget to create a good game with LLMs. They will definitely eat all your financial resources, because the cheap models are not capable enough for a complex 3D world. Yes you can get a bad Super Mario 2D clone, but not a complex 3D RPG.

Edit 2:
Okay there are such heavy issues in your arguments that I just need to speak about all the heavy ones.

The other thing that attracts a lot of scorn is people using it to create art "instead of employing an artist" but often times that's a false dichotomy. The choice is not AI art or human art, it's AI art or no art. The art in this context is simply not worth enough to warrant the cost of a human artist.

I was thinking this at first, too. But the reality shows that money counts more than anything and it replaces even work where artists are worse the art, but not the money. And even worse: those who actually do great art and employ people instead of LLM, spam the market with so much slop, that it becomes really really hard for human artists to survive, because they have to get found in all the spam. It was even a problem with human artists only. Now its twice as bad and it becomes worse in next years.

It even begins on creating your own webpage. LLM-generated webpages are optimized for search algorithms of Google and the others. Human made pages are not, even if they do at least some adjustments (as replacing a picture that Google does not like). People who hate LLMs already start to slopify their homepages, just to get back into the Google charts. Otherwise they would lose their entire business. Luis Rossmann was speaking about it and I know people in private how they lost their businesses after not being listed in top 10 any longer.

But I'm not entirely sure that the way AI has looked at art and uses it to generate more is all that different from the popularity of anime-style games.

It just doesn't matter how it is generated. The issue is the stealing in first place. They even steal from free software and knowledge projects. Free licenses do not mean they are allowed to use it the way they do. Let's speak about Creative Commons licenses.
CC0 -> Can be used for training without restrictions.
CC-BY -> There have to be a list of authors or sources next to the model which contains all original authors of the work which is used and training would be okay. To my knowledge not a single company of the big models is doing so. A violation of licenses.
CC-BY-SA -> This license forces everything trained and generated by NN to publish it with CC-BY-SA again. You know the reality, license is violated (including the issue from CC-BY).
CC-BY-ND -> No free license, but close. Even with this you can train your model, but you are not allowed to make money with it. Even advertising beside the product is forbidden. Big Tech would never use this license if they want to make money in a legal way. Yet this kind of licensed arts are abused and part of the big models.

Most companies do not care what data they are using. They just take everything they can get. They even know it is stealing as the example with NVidia shows. And at the end US laws seem to make everything "non copyright able", which leads to even further violation of licenses like CC-BY-SA or GPL code. And both are made to protect the art/software staying free for ever and to not getting prioritized.

they're using 150 watts

No, they own a PC that can run 150W or more. My PC is a gaming/workstation PC with 350W graphics card, but idling the graphics card eats 5-10W, 15W with some activity. Same with CPU, eating 3-5W idling. Depending on parameters a generator is using the full 300W power of the graphics card (don't want it to waste much power for little gain) for a very long time.

But even if it becomes more efficient. A human has only a limited amount of capacity. As often in tech, there is a paradoxon where more efficiency leads to more energy consumption. And NN can not just shrink the time of generating a picture from hours to seconds, so you probably generate 100 images in the same time someone draws a picture on PC and the 100 images require strong PCs (you don't want to lose too much quality compared to a real artist, so you are taking a better model) which runs on 100% usage when generating. The artist on the other hand draws the image most of the time with idling CPU, except the rendering tasks where a filter becomes applied etc.
Oh I told about 100 images? Why not 1000? The work is easily parallelize-able and can be automated. You even can use a NN to generate 1000 images and another LLM to filter the images to only get the 20 probably most useful where you can chose the one you need. Just showing where it could go if we follow the way big tech wants to lead us. NVidia is speaking about producing computers for LLMs and money, not for humans. So the goal is clear: make as much money as possible and let LLMs do jobs for the sake of doing something to generate endless growth of money. There is not even a real benefit behind this vision other than money.

Last edited by PlayingOnLinuxphone on 9 Jun 2026 at 4:31 pm UTC

Quoting: spymastermatt

Quoting: tuubiIgnoring any and all other concerns: As long as they don't then imagine that they've actually written a story, and try to claim copyright to the result, I guess it would be fine? I'm sure it feels rewarding for them to get that story out. They haven't become an author, more like an amateur editor, but still, good for them.

I guess this is where we differ. For me an author is the person that came up with the idea for the story, plot, characters, arcs, scenes, and development. So if the person (A) is doing all of that and describing them in the best detail they can to someone (B) who writes the actual words (be they computer or human) then in my mind A is still the author.

I don't think you can find a dictionary definition of "author"—at least in the literary context—that doesn't explicitly refer to the act of writing. Consider writers of non-fiction: They don't come up with stories, plots, characters, arcs etc. but they are authors. They might employ researchers to collect information to base their work on, and they might not even come up with the original idea, but the author is still the one actually writing the "story".

Ghostwriters exist, but they explicitly give up their rights to their work in exchange for money. Doesn't make the one paying for their work an author, in my opinion, even if legally they are just that. For example, I don't believe the current president of the United States ever authored a single book, but he sure has his name on the covers of several. Still more ethical than producing the books using an LLM, in my opinion. At least the original writer gets paid.

Quoting: spymastermatt

Quoting: tuubiI think the situation would be different if they'd actually written the whole story to the best of their ability, and used these tools to give them suggestions on how to make it better. A lot more effort, but there's a reason even established authors work for months, sometimes years, to write a novel, even with getting help and feedback from professional editors and readers. Most of us haven't got the talent or the patience for that, but that doesn't make it okay to just outsource the bulk of the effort.

That's a fair point and I guess then it comes down to how much of each. My friend (who is not actually considering using AI at this moment I should point out) has written several scenes of the book himself. If he uses an AI to help fill in the gaps between the scenes, is there a moment for you when it becomes not his work, or is it simply if he uses the AI at all? Genuinely interested in your opinion here, I appreciate from your later comment that my tone has not reflected that.

Seems simple enough: The parts he wrote himself are his, while the bits in the gaps are not. (Of course you didn't mean it so literally, but you know what I mean.)

Quoting: spymastermatt

Quoting: spymastermattI guess my analogy in this case would be is the songwriter who writes clever lyrics, less of a songwriter because they can't sing it themselves? Should we scorn them for using a singer instead of just singing it themselves like a real artist

Quoting: tuubiThe analogy doesn't work. Why wouldn't they be called a songwriter after writing a song? They'll get their share of the royalties. They just can't be called a singer. Just like your friend isn't an author after prompting and editing a book.

I guess then the problem for me is that we don't yet have a view or word to separate the ideasmith from the author in the way we do the songwriter and singer.

Sure. If you think the title matters. Although I'm quite sure I've read a ton of books where the author thanks other people for giving them the idea for or helping them develop the story.

All I'm saying is that your friend is literally (pun not intended) not the author of any text he didn't write. Of course, I'm not sure that matters, unless he wants to make money off of the result. I don't think he should be able to sell a story he didn't write and claim to be the author, but maybe that's just me.

Quoting: spymastermatt

Quoting: spymastermattAgain, we've taken the idea that AI can make a book given a one sentence prompt and blanket swept that that must therefore be the only way people use it

Quoting: tuubiNo, what you've done is ascribe a lack of understanding to people with valid criticisms of the technology and how it is being used in the real world. Sure, some people react with their gut, but critics do include a lot of people who are actually in the know, but crucially not invested.

I apologise that this is how I have come across, I was intending to question what feels like gut reactions but I appreciate that has come across as assuming people don't understand.
I was trying to separate PurpleLibraryGuy's suggestion of

Quoting: Purple Library Guyif someone tells a large language model "Write me a book about this" and thinks it's theirs because they had an idea (which apparently does happen) . . . that's incredibly stupid.

from the idea of people using AI to fill in gaps rather than write the whole story. From my perspective there's a big range between Wrote the book entirely myself -> Had the AI spell and grammer check -> Had the AI fix up some scenes to make them read better -> ... all the way through to "Told the AI to write a sci-fi book about an alien who lives on earth" (which I 100% agree removes any creative direction from the human)

And yes, a lot of the reactions to every announcement of AI use in any form feel to me like
"100% AI = bad, therefore 1% AI = bad"

Incidentally, you might find a "creative director" in the credits of a game. They direct the developers and artists to achieve a coherent vision, but do not get to claim credit for their efforts.

I'm not sure if you'd consider this anecdote a good illustration of my point, but I'll tell it anyway: Years ago, I commissioned a skilled carpenter to make a piece of furniture according to my specifications. I designed the thing, I made nice schematics and drawings, with precise measurements and details for how I want the parts cut and joined and everything, but they're still the one who actually made it in the end. I don't mind that I don't get to call myself a carpenter now. Not that I would call myself a furniture designer either, at least with a straight face, but that's beside the point.

All that said, I fully understand and empathise with your viewpoint and that of your friend. The world is full of very creative people with their heads full of great ideas, but most of them are not able—for one reason or another—to make anything of them. I've always been an avid reader and gamer myself, and I've obviously come up with (and sometimes even written down in detail) my share of story and game ideas that'll never see the light of day. I suppose that's a shame. Assuming those ideas were ever any good. 😅

Last edited by tuubi on 9 Jun 2026 at 3:32 pm UTC

Quoting: pbYeah, the AI is incapable of replacing artists, but it's capable of chewing up all the artists' works and spitting out something more or less acceptable by an executive. So yeah, it won't replace artists but it will be used (by CEOs) as a replacement for a junior illustrator, who with some work, persistence and money (to support day-to-day living) might have grown to be an artist, something that will now be a bit harder...

Personally, I think the only thing AI has any use for is for brainstorming ideas, and maybe creating textures that would've been easy to generate procedurally anyway. They won't be replacing talented junior artists. If you're the team lead for a project, and your standard is to churn out a product that's only just "more or less acceptable", your audience is going to notice, and you'll pay for it in the end in a way that you didn't intend to. They don't call it "AI slop" for no reason.

Quoting: spymastermattI saw a fantastic sculpture in Birmingham UK made out of knives from an amnesty. It must have taken hours to put together, but that's not what made it impressive. It was good art (to me) because of what it represented.

I'm only just touching on the surface when I talk about appreciating the time and effort it takes to produce something worth looking at. Like you've mentioned here, part of what gives a piece of art meaning is what its creator meant to represent. An artist doesn't necessarily have to have the technical skill to create something that's very moving and unique - it just has to have passion. Cavemen painted on walls 1000s of years ago to communicate how his day to day life made him feel, and maybe as a record of history that he wants echoed through time. Beethoven composed Fur Elise to express how he felt about a woman that he loved. Michaelangelo sculpted the Pieta to show his interpretation of what a mother must feel to have her son die in her arms. Machines can't imbue art with any meaning, and at the end of the day, that's really what it's about.

Last edited by Slaxer on 9 Jun 2026 at 4:12 pm UTC

Did a second Edit. Uff sorry for that wall of text, but arguments require some room.

Quoting: spymastermattFor me an author is the person that came up with the idea for the story, plot, characters, arcs, scenes, and development. So if the person (A) is doing all of that and describing them in the best detail they can to someone (B) who writes the actual words (be they computer or human) then in my mind A is still the author.

Quoting: spymastermattI guess then the problem for me is that we don't yet have a view or word to separate the ideasmith from the author in the way we do the songwriter and singer.

It sounds as though you don't really understand the work of artistic creation. I don't mean that as an insult; many people don't know the work that goes into something like writing a piece of fiction. (It's also possible that you don't understand Generative AI technology, though I find that less likely given your participation in this conversation.)

Ideas are the easiest part. "Good" ideas are just ideas that have been executed well. Ideas for stories are so common that there is a saying about everyone having a book in them. If you ask a verbal toddler for a story idea, they'll have one. That's how easy ideas are.

In written fiction, every single word represents a choice - at least one - that the author is making in service of the overall piece. Often, a writer will try one choice only to discard it in a later revision. I sometimes write entire scenes only to discard them later on. That doesn't make them useless; I learn something about the story as I write and revise, even as I discard the majority of the written work. All of the choices, be they in a major scene or "the gaps between the scenes," are important for the way an author expresses their work and the way a reader encounters the story.

Writing in a style and tone that fits the author and the idea (plot, character, theme, etc.) and refining that combination over multiple revisions **is the work**, and it is through that process that the depth of character and theme are realized. Purple Library Guy mentioned an aspect of the process that I quite like: at times, things happen in a scene or in a piece of dialogue that the author did not expect when sitting down to write. When that happens, the author discovers something about the piece, about the character(s), and maybe (if they have some introspection) about themself. This will change the way the author approaches the next scene, for instance, and may change the way the author writes the entire piece in the next revision. This discovery (and self-discovery) is what comes from doing the work, and is part of what makes that work into art.

This may be why we don't have a standard word for a person who merely comes up with ideas (maybe "collaborator" is the best word, though really, people generally get no more than a hat tip in the acknowledgements section for solving major plot problems). Ideas are easy. Writing, for most people, is hard. It gets easier with practice, but even then it is a matter of wringing the best version of a thought from your brain and presenting it for your reader in the most meaningful way, and then doing it again and again and again. Then getting up the next day and doing it again and again and again, sometimes discarding everything you did the day before. This constant chipping away in service of the vision is the creation of art. That's the process.

Quoting: spymastermattFrom my perspective there's a big range between Wrote the book entirely myself -> Had the AI spell and grammer check -> Had the AI fix up some scenes to make them read better -> ... all the way through to "Told the AI to write a sci-fi book about an alien who lives on earth" (which I 100% agree removes any creative direction from the human)

There are shortcuts to the writing process, and often people feel differently about products that come via those shortcuts. We have the concept of ghostwriters, for instance, which is a bit like what you're describing in the first quotation way up above: Person A has an idea and Ghostwriter B writes it for them. This is generally a close collaboration between two people, with the ghostwriter often trying to capture as much of Person A's voice and emotion as they can. Person A usually gets listed as the author, but nobody would compliment Person A on their writing. They didn't write it. In literary circles, Person A would get zero cred for this kind of work and Ghostwriter B would typically only mention it insofar as saying they got paid to write something.

If someone uses another person's fully realized ideas (characters and world, for example), we call it fan fiction. Many people (perhaps most, fanfic communities aside) would consider that work "lesser than" the original, regardless of the quality of the fan fiction. In literary circles, fan fiction authors typically get no cred for this kind of work. Ideas are easy; use your own. At the same time, nobody considers J.R.R. Tolkien to be the author of any of the Legolas fan fiction (even though the characters and world are his); coming up with the ideas isn't enough on its own to make you the author of a piece of work.

To address your specifically mentioned shortcuts:

If someone were to use Gen-AI to spell-check their fiction, I would wonder why they needed AI to do it; we have spell-checkers that don't rely on Gen-AI (and thus don't fall afoul of the various compelling energy-use arguments against Gen-AI). Still, spell-check is a relatively minor change to writing that usually doesn't affect meaning; if someone wants all of their work to be spelled in a standard way, fine. I'd recommend being careful about dialogue and other parts of writing that are heavily affected by character voice.

If they were to use Gen-AI to do a grammar-check, I would start to feel uncomfortable. Grammarly, for instance, claims to correct one's grammar, but it actually re-writes text to conform to the average writing in its corpus. It does not take into account the emotional content or other intention behind a piece of writing. It does not understand that there are situations that call for an abnormal sentence structure or a sentence fragment. It can't hear the tone and rhythm of language and other factors that constitute an author's voice. If you're writing dialogue, it doesn't understand that people often speak in ways that are not grammatically correct. Gen-AI tools can be prompted to try to account for these weaknesses, but you're still going to get the average of whatever that model can generate according to your prompts rather than work that fits the specific emotion and meaning that you want. If someone were to use Gen-AI for a grammar-check, I would want them to go through and re-write the work afterward to correct the voice - at which point, I'm again forced to wonder why they chose to use Gen-AI. A human is far more likely to pick up the emotional content and voice (and understand the author's intent); just use a human editor and go through the process of writing, revising, and refining your work until you're comfortable with the grammar. Your writing will improve over time.

The next step in your spectrum is having Gen-AI "fix up some scenes to make them read better," which is begging the question (in the philosophical sense) of what Gen-AI can do. It can't write a section to improve someone's voice; it doesn't know an author's voice better than they do themself. If someone doesn't like their writing voice, they should work on it; Gen-AI can't do that for them, either. I guess I'm not entirely sure what is meant by "make [the scenes] read better," because Gen-AI simply can't do that. It can make things read *average*, or even the average of a specific prompted style, but neither of these things is "better" than what a writer can achieve by iterating on their work over time.

The most common conclusion of that saying about everyone having a book in them is something like "and that's where most of them should stay."* This is generally taken to mean that most people shouldn't try to write a book because writing well is hard work. I agree that writing well is hard work; I disagree with the idea that most people shouldn't attempt to do it. I think that the more people try to capture their thoughts and emotions in their own words, the more likely they are to appreciate the work of others who try to do so as well.

*This is often attributed to Christopher Hitchens, though I've never seen a record of his actually saying it.