• Most new users don't bother reading our rules. Here's the one that is ignored almost immediately upon signup: DO NOT ASK FOR FANEDIT LINKS PUBLICLY. First, read the FAQ. Seriously. What you want is there. You can also send a message to the editor. If that doesn't work THEN post in the Trade & Request forum. Anywhere else and it will be deleted and an infraction will be issued.
  • If this is your first time here please read our FAQ and Rules pages. They have some useful information that will get us all off on the right foot, especially our Own the Source rule. If you do not understand any of these rules send a private message to one of our staff for further details.
  • Please read our Rules & Guidelines

    Read BEFORE posting Trades & Request

Using AI for cover art

It's not what I'd use for every edit I make, but there's definitely ideas for edits I have (like these ones) that it suits perfectly.

Yeah I've had a couple where I haven't been able to generate anything useful related to the concept I wanted - Resident Evil was a hard one because I specifically wanted something with Alice and Red Queen, and it would give me little anime girls instead of the Red Queen when I was trying to get her hologram character. And Alice is such a generic concept that keywords specifically for her are difficult.


I eventually went a different direction and made a very old-school style by hand:


Another one it seems to have difficulty with are lesser known buildings. I wanted 2d or line art resembling a bank Eero Saarinen made in Columbus Ohio and I gave it specific enough prompts that it should have been able to identify the building but it wouldn't make anything even close (the bank features heavily in the movie Columbus).

It made some interesting art of Columbus Ohio, just not what I was actually looking for...
 
a lot of these images have elements that experienced artists will be able to pick out right away. I think this can be problematic if people who are against AI see your cover and just turn their back on your work based on that. I used AI for the first version of my Kang cover, as I couldn't find stock images for what I wanted, so I generated an antman figure and a small army. I regretted it in time because I knew I would lose some potential viewers, so I spent some time and found images that suited the purpose. I don't think AI images are terrible as placeholders, but if you can, I strongly encourage getting a human to replace the AI elements. Art is communication and AI doesn't really communicate anything. It's the human element that we need.
That LOTR image for example, I'm glad the floating mountains were cut out, although they are still partly visible. I think the image you presented communicates, thanks to your layout, but the whole AI makes me uneasy from the off.
That may be true for some, but it doesn't seem like the niche that is this community are expressing a lot of concern regarding using Ai as a tool and resource. We're vandalizing art everyday here with fanedits. TM2YC did much more than simply enter a prompt and slap a title on it. There is a ton of human in their piece. I personally don't see any issue with it in this manner.
 
I was messing around with the aforementioned Columbus prompts a bit more just to see if I could get it to generate Eero Saarinen's iconic pointy church instead, and while I didn't succeed, some of the results were pretty neat. I went with a line art style for this set (prompt: "ohio Eero Saarinen pointy dark church line art")

I don't know if I'm going to use this for the cover since it's not a real building, but it does capture the aesthetic I was looking for as a cover:
 
If you have a current Adobe subscription, you can use their generative AI. It was trained on Adobe Stock, a library of photos and videos Adobe paid real artists for. If paying or crediting the artist that the AI models were trained on is a serious concern for the kind of usage this thread is demonstrating, then Adobe may have the most ethical (and expensive) solution currently.
 
If you have a current Adobe subscription, you can use their generative AI.
I'll have to play around with that, I took a quick glance at the Adobe Firefly site and it looks like they give you a lot more control over the different elements in the image too which seems like it will make the process quicker and more adaptable than just regenerating the prompt and hoping for a better one.
 
I'll have to play around with that, I took a quick glance at the Adobe Firefly site and it looks like they give you a lot more control over the different elements in the image too which seems like it will make the process quicker and more adaptable than just regenerating the prompt and hoping for a better one.
I've been using it for a while. I "created" my Star Trek: Regenerations Blu-ray cover by using the outpainting feature on some official art which is very narrow. Not the most wildly creative use of AI, but it did a good job.

And yes, the original artist was credited.
 
Yeah I've had a couple where I haven't been able to generate anything useful related to the concept I wanted
I think the AI finds it easier to generate things it's seen more pictures of.
A superhero stood on a rooftop is easy because it's been done a million times, but if you want two superheroes together, there's way fewer images it can draw from, so it looks at individual pictures and then gets confused about what elements belong to what character.

After many attempts, I made this poster by generating Green arrow on a city street and Deathstroke separately and then combining them.
Poster.jpg
If I tried to generate an image of them together fighting for example, it would always give Green Arrow a mask, or give Deathstroke a hood and a quiver of arrows.

I think you might be able to do your Resident evil idea by separately generating, "red hologram of small girl in a dress", "city street overrun with zombies" and "Kick-ass woman with short hair and two guns stood on top of a car doing a kick"
And then combining the three images together.

most of my edit ideas don't really suit an AI generated poster, but I still might use it to create smaller elements of posters, like a background or something from the movie shot in a way that isn't in the movie, and then combine those elements with a real picture of the main characters for example.
 
If you have a current Adobe subscription, you can use their generative AI. It was trained on Adobe Stock, a library of photos and videos Adobe paid real artists for. If paying or crediting the artist that the AI models were trained on is a serious concern for the kind of usage this thread is demonstrating, then Adobe may have the most ethical (and expensive) solution currently.
Yeah, it's interesting and I agree. And I also agree with @tremault . This is a fascinating and volatile topic. Unfortunately, just because artwork is in Adobe Stock doesn't mean all artists consent them being used for generative AI, where artists names show up on datasets for generators to mimic their style.

This has been a continuing point of contention where some artists felt blind-sided by this move. There's still a lot of work to achieve ethically AI stock images where artists permit and are properly compensated for this specific end use.

Now, to use it or not use it as coverart? I'd treat it the same way as fan-editing. We're all here. I think the least anyone can do is find a way to credit the original artists and/or prompts somewhere (like even as small as a caption, maybe even separate from the AI generated piece itself) for the images used. And absolutely NO GAINING MONEY from selling that kind of art- just as anyone would approach sharing their fan-edits, duh. That's the very least anyone could do, I think. It's the same way I've seen folks approach their posters. I've seen incredible fan-art done by artists who aren't the fan-editor, and I think that's perfectly cool as long as artists are credited (and no selling, duh).

STORY TIME (please don't read ahead, unless you want to)

I don't personally use generative AI in visual artwork. I'm very particular, and I've found there's fulfillment in sucking and achieving- from learning and doing what you want to see happen. I don't think I'll ever be a graphic designer in my life. But I remember in grade school that I liked to draw, and I drew for my classmates. There was this kid who wanted something drawn (I forgot what it was, it was like Batman and some pictures of our gym teacher with Hellman's Mayonaise logo in his chest- something stupid like that). I drew it for him, and then he would go around telling other kids from other classes that he drew it. And I'm like, ooh huh? No one believed him. And no one liked that kid. He had no friends. A loser, as they say. But it was whatever. Just a kid and his infantile ways- cringe and smells of some people Artisdead, you know? Bad vibes. Makes you feel icky.

Anyway, what was I saying? Oh yeah, there's fullfillment in learning to do what you'd like to see. It just sucks that Photoshop and similar tools are criminally expensive. I said this loudly before- You are human. You don't just get fed images of stuff and compile to make art. No matter what AI bros want you to believe, "AI" doesn't learn the way humans learn. You made art the other day because you met a very cute barista, and it made you feel you wanted to draw something. You know what you drew? A flower! You didn't get fed images of flowers that you wanted to draw. You thought of that cute barista and how they made you feel. That's art, it's your flower, your art (it sucks but that's okay). I went for a beautiful walk and thought to myself, you know what? I want to sketch some hardcore, pornographic giantess-babe drawings. That walk, it made me feel something. Inspiration can come from anywhere, not just images of hot giantess babes fed into you. When I draw a banana, I don't think of pictures of bananas. It's not just input-output. I'm drawing a banana because I just ate one, maybe. Taste, hearing, feeling. AI can't do that. And when AI eventually does, it'll be the AI that gets credit (because it lived, it got its heartbroken, started a family, whatever)- not you, the AI-bro, bad-smelling-kid-from-grade-school who wrote the prompts. Yeah, when you draw, the drawing will suck at first. But it's so good because it's yours, it's your idea executed.

Look at this.
does-anyone-remember-this-fan-poster-for-star-wars-episode-v0-49eyuf8jsryb1.jpg


Look at that. You think that's awful? No! It's gorgeous (it's actually pretty good). And you know what? That's not mine. I don't know who made this, but let me tell you, that person is probably doing great now, and they're proud of it. And I'm proud of them too.

Anyway, I thought I'd throw my two cents here since I'm just finding out about this AI thread. I've been working on a Lion King edit. The whole thing uses an AI voice narration (and it's a complete mess so far). There's no other way unless I kidnap Sir David Attenborough myself. Have a good evening folks, it's late.
 
I think the AI finds it easier to generate things it's seen more pictures of.
A superhero stood on a rooftop is easy because it's been done a million times, but if you want two superheroes together, there's way fewer images it can draw from, so it looks at individual pictures and then gets confused about what elements belong to what character.

After many attempts, I made this poster by generating Green arrow on a city street and Deathstroke separately and then combining them.
Poster.jpg
If I tried to generate an image of them together fighting for example, it would always give Green Arrow a mask, or give Deathstroke a hood and a quiver of arrows.

I think you might be able to do your Resident evil idea by separately generating, "red hologram of small girl in a dress", "city street overrun with zombies" and "Kick-ass woman with short hair and two guns stood on top of a car doing a kick"
And then combining the three images together.

most of my edit ideas don't really suit an AI generated poster, but I still might use it to create smaller elements of posters, like a background or something from the movie shot in a way that isn't in the movie, and then combine those elements with a real picture of the main characters for example.

I had a go with another service called Picsart to make this look more like a comic book cover:
Comic-Poster.jpg
I'm not sure if I like it better than the "Realistic" version, but the title looks really cool in this one.
 
So I was curious what the AI would do with the prompt "westworld", so I didn't give it anything else, and it didn't reject it. Instead I got these rather... interesting depictions:

I liked the idea of the metal cowboy and horse motif so I honed the prompt in that direction with a variety of different prompts like:
"scifi old west chrome horse chrome cowboy robot with cowboy hat in desert at sunset"

After a few iterations I decided I liked the female style better, and a few tests of that gave a lot of really janky hair so I targeted models without hair:
"scifi old west chrome hairless female cowboy robot with cowboy hat on chrome horse in desert with buttes at sunset"

And then just to see what it looked like with a bit less shine to the metal:
"scifi old west matte chrome hairless female cowboy robot with bandana and cowboy hat on matte chrome horse in desert with buttes at sunset"

That's only a small selection of the most interesting, generated 50+ total. Final cover:
 
Yeah, it's interesting and I agree. And I also agree with @tremault . This is a fascinating and volatile topic. Unfortunately, just because artwork is in Adobe Stock doesn't mean all artists consent them being used for generative AI, where artists names show up on datasets for generators to mimic their style.

This has been a continuing point of contention where some artists felt blind-sided by this move. There's still a lot of work to achieve ethically AI stock images where artists permit and are properly compensated for this specific end use.

Now, to use it or not use it as coverart? I'd treat it the same way as fan-editing. We're all here. I think the least anyone can do is find a way to credit the original artists and/or prompts somewhere (like even as small as a caption, maybe even separate from the AI generated piece itself) for the images used. And absolutely NO GAINING MONEY from selling that kind of art- just as anyone would approach sharing their fan-edits, duh. That's the very least anyone could do, I think. It's the same way I've seen folks approach their posters. I've seen incredible fan-art done by artists who aren't the fan-editor, and I think that's perfectly cool as long as artists are credited (and no selling, duh).

STORY TIME (please don't read ahead, unless you want to)

I don't personally use generative AI in visual artwork. I'm very particular, and I've found there's fulfillment in sucking and achieving- from learning and doing what you want to see happen. I don't think I'll ever be a graphic designer in my life. But I remember in grade school that I liked to draw, and I drew for my classmates. There was this kid who wanted something drawn (I forgot what it was, it was like Batman and some pictures of our gym teacher with Hellman's Mayonaise logo in his chest- something stupid like that). I drew it for him, and then he would go around telling other kids from other classes that he drew it. And I'm like, ooh huh? No one believed him. And no one liked that kid. He had no friends. A loser, as they say. But it was whatever. Just a kid and his infantile ways- cringe and smells of some people Artisdead, you know? Bad vibes. Makes you feel icky.

Anyway, what was I saying? Oh yeah, there's fullfillment in learning to do what you'd like to see. It just sucks that Photoshop and similar tools are criminally expensive. I said this loudly before- You are human. You don't just get fed images of stuff and compile to make art. No matter what AI bros want you to believe, "AI" doesn't learn the way humans learn. You made art the other day because you met a very cute barista, and it made you feel you wanted to draw something. You know what you drew? A flower! You didn't get fed images of flowers that you wanted to draw. You thought of that cute barista and how they made you feel. That's art, it's your flower, your art (it sucks but that's okay). I went for a beautiful walk and thought to myself, you know what? I want to sketch some hardcore, pornographic giantess-babe drawings. That walk, it made me feel something. Inspiration can come from anywhere, not just images of hot giantess babes fed into you. When I draw a banana, I don't think of pictures of bananas. It's not just input-output. I'm drawing a banana because I just ate one, maybe. Taste, hearing, feeling. AI can't do that. And when AI eventually does, it'll be the AI that gets credit (because it lived, it got its heartbroken, started a family, whatever)- not you, the AI-bro, bad-smelling-kid-from-grade-school who wrote the prompts. Yeah, when you draw, the drawing will suck at first. But it's so good because it's yours, it's your idea executed.

Look at this.
does-anyone-remember-this-fan-poster-for-star-wars-episode-v0-49eyuf8jsryb1.jpg


Look at that. You think that's awful? No! It's gorgeous (it's actually pretty good). And you know what? That's not mine. I don't know who made this, but let me tell you, that person is probably doing great now, and they're proud of it. And I'm proud of them too.

Anyway, I thought I'd throw my two cents here since I'm just finding out about this AI thread. I've been working on a Lion King edit. The whole thing uses an AI voice narration (and it's a complete mess so far). There's no other way unless I kidnap Sir David Attenborough myself. Have a good evening folks, it's late.
I really enjoyed your post here.
it's given me an idea, I'm going to start a new topic about how I put together my covers and what the sources I used are.
 
I really enjoyed your post here.
it's given me an idea, I'm going to start a new topic about how I put together my covers and what the sources I used are.

Cool idea!

Edit: I can also contribute to that thread with the opportunity to share what layers and sources I’ve used as well in my covers!
 
Last edited:
Okay one more for today.. Hunger Games. This one was fairly easy to generate images for, but took a lot of effort to actually create something decent because AI really sucks with details like fingers and bow strings, arrow position, feather orientation, etc. If you look at the two source images I cobbled together for the final cover you'll see a myriad of changes to blend them properly and correct all the AI idiosyncrasies.

The results do seem like the AI combined the Archer character with the Hunger Games girl though, you see a lot of hoods and masks, as well as a lot of Asian girls that might have come from cosplayers or some other similar archer character.

Prompt: "survival archer girl gold mockingjay on fire"
Final Cover:
 
Yeah I've had a couple where I haven't been able to generate anything useful related to the concept I wanted - Resident Evil was a hard one because I specifically wanted something with Alice and Red Queen, and it would give me little anime girls instead of the Red Queen when I was trying to get her hologram character. And Alice is such a generic concept that keywords specifically for her are difficult.

I've decided to give Resident Evil a shot:



----

And yeah, AI sucks hard when it comes to bows and arrows.
 
Last edited:
I've decided to give Resident Evil a shot:



----

And yeah, AI sucks hard when it comes to bows and arrows.
Some of these look a lot like the actor. What prompt did you use to describe her?
 
Well, "female Jovovich" worked fine at around 40% of the results I guess, then some details about hair and clothes :>
 
Well, "female Jovovich" worked fine at around 40% of the results I guess, then some details about hair and clothes :>

Very nice, didn't realize you could use part of an actor name to sneak through the filter
 
Back
Top Bottom