Hey all,
I know the topic of AI is a rough one for a lot of creatives. It's threatening the industry I spent nearly a decade in, so I absolutely understand the feelings around it. My own two bits are that it has potential as a tool and is fine to use it to make art for personal projects or things which make no money (like a D&D game with friends), but the moment it's used to make money it needs much more care and regulation, especially considering how it references others' art, etc etc.
So, as the title asks, how do people feel about making art for covers/chapters using AI tools? I was playing around with some prompts for my story The Outlaw, The Star and the Big, Big Sky and it turned out some really neat ones which have a unique vibe and quality to them. I was considering swapping out the current cover art for one of these, or inserting it into the story. But I wanted to get a feel for the community's opinion first.
To make it clear: I would only use this art as a point of fascination, a "hey, check out what it could look like" sort of thing. I'd never claim it as my own, and I am no AI apologist. I like to think of myself as an artist too (I'm working on a graphic novel at the moment) so I fully understand that people have a strong dislike for AI art. I really am genuinely curious and interested to hear what the community here thinks.
Here's one such image I generated. I'd want to do some editing before using anything (It sure can't handle text very well, haha), but I just thought it was neat.

7968291
Yeah, I agree with you like. Like for personal things, soundboarding, or to just fuck around, that's one thing, or if say, you're an indie game dev who has little to no excess funds to pay for handmade art or models or VA work aside from friends, then that's understandable.

7968293
Totally. I think people should be clear that they used AI and aren't claiming to have drawn it on their own, in any case.

7968291
Personally I have no problem with it, I’ve found it to be a pretty useful tool. I can’t draw anything better looking than a stick figure, so my options for cover art are either paying for a commission or using generic images. I used Dall-E to generate some art for my story Night of the Autumn Moon, and I really liked what it came up with. Also, I had to do some photoshopping to get it to look right, so I consider it AI-assisted rather than generated.

It’s a nice middle ground if you want nice looking original cover art, but don’t want to spend money.

7968291
Don't personally have an issue with it, but I'm also interested in AI, so take that as you will.

Most AI picture generation can't handle text very well. It's something they're constantly working on. What you can do is after the picture is generated, go and edit it and manually put in what you want it to say, and then inpaint over the area you edited, to get it to blend in more with the rest of the picture.

Same goes for fixing up other issues with the image, really.

--Sweetie Belle

7968300
There's also the grey area of using AI to make a base and then editing and tweaking it oneself, IE some artists I know use AI to make the base, the bare necessity stuff, than adding their own flair

7968302
That cover art looks really good! And if I had all the money in the world, I'd gladly toss it around at artists. Heaven knows there's been times where even I've offered to do commissions and haven't been able to get any takers.

7968312
AI is here to stay, regardless of how people feel. I worked with a small startup company for a couple months recently which was working with AI, and yeah, companies are heavily investing in it. It's up to us to figure out how to use it and keep up before the boat leaves.
And honestly, I'm pretty impressed with how well it handles text as is! haha

7968362
Or even vice versa, doing the sketch and then having the AI color and render the rest. It's going to be an interesting couple years for artists as people figure this all out.

Good to know that the general sentiment seems to be fairly close to my own.

7968291

Being a 2d and aspiring 3d artist, I am of mix opinions on on it on AI art. For 2d stuff I feel, while is can look impressive at first glance, until you look a little closer to de details. You also had to understand that most art is not obtained through very legal means, but also that it can't be inspired by any other media like literature, news, music in generating so it's. The basic principle of a AI art is that it deconstructs and image and backwards rebuilt images enough so that it can't be claimed to just be a copy of an original, Tech bros exploiting a legal grey zone that the law is struggling to catch-up with. And you are still paying to use those program's tools that are based of datasets of artists they didn't ask permission to use, and wouldn't be able to generate any of it without obtaining them through dubious means.

AI art is also becoming increasingly associated derivative and cheapness, only creating what has been done before, along with cons, (NFT) and money skims, the infamous Willywanka factory tour is a good example of that. The issue with AI art at the moment is that the people who use it don't understand style, composition, or good taste, and doesn't realize what they think looks good will be shared with by others people. And when your audience realize it's AI you can't unseen it and it will be mock endlessly for it.

On the 3d side it's a little more blurry, with how time consuming, and boring it takes to create something and you need a find ways to speed it up the process like generating texture, realistic of stylized or creating random MPC's and not everyone is good at creating thing procedurally with a node base system. There is also to add that not all modelers or 3d sculptors are trained artists and you some time just need a quick idea to run off on to jump into a project, the rest is done by hand. Computer assisted in generate repetitive props has been a around a long time in animation and games for stuff that you would never bother or able to look really closely.

At the moment the issue of using AI in the industry isn't really the treat but merely the pretext to devaluate the work of the artist to keep their salaries low and a pretext to outsource the jobs overseas where labor is cheaper and unions less organized. And it's not really explained the process of generating art or text, but we have learned that a lot of data scraping filtering is done in third world countries payed 2 USD an hours to look at the most horrifying thing that the deep dark former of the web. An interesting second hand anecdote I was told of a friend who words in the game industry was told that they weren't allow to generate anything with frames in it because they kept generating burning human figures, I will let you guess why, as it's an understandably very touchy subject.

Fortunately there has been countermeasures to poison AI samples and it's now a constant arms race between protecting art and extracting, making it that much harder for AI to analyze.

On extra note, there is also talks from Tech bros about AI generated music, they have been suspiciously quick to point out they would take non copyright free songs to generate it. And when you consider how aggressive the record companies are in protecting in protecting their intellectual properties in the past, it wouldn't come as a surprise that AI companies would be sewed to oblivion by the record companies, and cast AI generated writing and art under a less favorable light to in associating it as stealing.

On the more positive use of AI generated art is this 2d animated project from an old early 200's cartoons where they use inhouse data sets to convert 3d animation to look more like the original show's 2d animation style.

7968415
This is a very detailed breakdown of the uses and pitfalls of AI. I agree, especially in the potential for misuse and legal/moral grey areas which arise. I think, unfortunately, the first jobs which will be made obsolete by AI will be concept artists, particularly environment concept artists. Anyone who's played around with AI doesn't have to try very hard to realize just how cost effective using Ai vs an artist to produce concept art is.

I was actually in the animation industry for close to a decade so I relate a lot. Especially to the complexities of node based rendering, which is something I could never really figure out haha I think it's particularly threatening to "simple" animation like commercials and children's animation, and those are going to be threatened very soon, if they aren't already. It's going to be a very strange 5 years or so while this whole industry is shaken around, which is largely why I'm stepping away from it for a while.

7968426
Environment art is one of the most boring, repetitive and least glamorous part of the industry and a lot of it is coloring with existing models on an image, so not big lost there and tweaking with filters to get a slightly different look can create drastic different, but again the concern would be that it will become derivative. Concept art will always be needed to one degree or another to establish the design mood and style and people don't nesseceraly realize that a lot of concept art is basically photo bashing to save on time and creativity and artistic freedom is very narrow. It's just that you shoulder off a lot of the boring stuff off artist. it just means that they can people can create studios with less people needed, which is one of the reasons why indie studios are really starting to get popular now, if I can take it as a form of positive.

For nodes stuff I am still struggling to understand it. I do get the basic principle for textures for blender, and I can almost confidently say I can create my own textures now, but procedural stuff like random environments, effects, or simulations I am still scratching.

Again what is the biggest problem isn't AI itself, but corporations that are outsourcing the animation process and art even more then before, constant meddling in the creative process and studios constant turnovers losing it's talents and experienced people and replacing them with younger less experience artist that they have to train along with producing the animation and replacing them soon after anyway, and upper ups meddling, it's soul draining for the staff.

What did you do exactly in the industry?

7968444
I’ve only played a little with blender. Most of my time was spent in Maya and I learned new ways to hate programming thanks to it, haha.

And yeah, companies will abuse it if they are allowed. It’s why I think the next couple years will be critical for figuring this stuff out.

I started out as a layout artist and worked my way up to animation director on various shows throughout my time. It had its ups and downs but I learned a lot, including about myself and my own skill sets. Storytelling was always at the core of my passions though, which is why writing is always a core tenant of my hobbies.

7968463
Do you have a portfolio or gallery? I am curious to see your stuff. Here is my DeviantArt gallery and Youtube channel.

7968488
Your stuff looks good! lots of vibrant colours and high detail pieces. I'll have to take a closer look soon.
I'm not quite ready to share much yet, mainly because of a desire for privacy. But I will PM you.

7968384
Yeah, I've seen it divide artists. Some I know use it so they can do more art, especially as then they feel less of a slave to the art

7968647
I understand that feeling, and I have friends who can't draw who have expressed how nice it is to have a way to actually bring their ideas to reality through the use of AI.

7968831
Yeah, its a double egded sword

By the way, I ended up using AI to make a new cover art for my story The Outlaw, The Star and the Big, Big Sky. It was actually pretty fun to make a bunch of images, bring them into an art program and combine and tweak them all together to become what I want.

7969974
Looking good!

Generating multiple pictures and then combining them in an art program is definitely a good way to go, especially since you can always bring the picture back for more passes with image2image.

One other interesting option: There's a plugin you can download for Krita that will either set up ComfyUI for you or connect to an existing copy, and add a sidebar where you can generate ai art and then copy it into a layer, or you can use the Live option. In live mode, you type a description in, then start drawing, and it starts generating ai art based on a combination of what you are drawing and the description. Really neat stuff, and something I've been meaning to play with more...

--Sweetie Belle

7970075
Oh hey that’s neat. I just started using Krita recently. I’ll Have to look into that, it sounds really nifty!
And thanks! Yeah I think as the art community moves forward, getting good at fixing AI art will Be an important skill for artists to have in the workplace. Whether or not that’s a good thing, I can’t say. But to stay competitive, it’ll probably be necessary.

7969974 7970075
ArtBreeder

7968291
Do it. If people don't like it, then that's their problem, not yours.

I think A.I art is overrated.... i doubt A.I can draw some pony town characters to be different

as such

this is my Pony Town character and I drew this... digitally and you know what... I'm proud to be an artist.