The Pros and Cons of AI in the Comic Book Industry

[KurtW95]

I was wondering what other people were thinking about this topic. I understand that there are some people who are really worried that it could make their jobs go away and other who like it because it could allow them to more easily create comics. I think there is a nuanced view that might be the best way forward. First of all, I do not think AI should replace the jobs of people in comics. But I do think AI could be used to help people in comics. We know that a lot of programs are way more advanced than the free ones that everyone can use, which have been majorly nerfed. I think it could help editors in running the script and checking if there are any continuity errors. Simple ones and ones that people may not know if they haven't read a single back issue of hundreds. You could also ask if the script is writing a characters in character. And none of this would harm the jobs of editors, because they would still be doing what they do. But just using AI as a tool to help. It could also be used to help writers who can't draw well to make mock-up storyboards to help give the artists an ideas of what the writers are wanting.

[MajorHoy]

Sometimes it's hard to find signs of human intelligence in the comic book industry . . .

[ed2962]

I don't think these are bad ideas in theory, but Marvel hasn't cared about continuity since Joe Q took over. DC only cares every 10yrs until the next reset. As far as storyboards and page layouts...maybe? But alot writers who feel the need to do such things get by with stick figures and roughs because they want the artist to interpret things in that artist's own way. AI art might compromise that.

[MRP]

Until all the material used to train the AI can be properly sourced with transparency, all the sources are cited and properly compensated, there are no pros to using AI in comics. And since the turn to AI is a cost-cutting measure so they don't have to pay people with the ability to do these things a fair wage, they will never compensate those who create the material that is used to train the AI either. It's another exploitive tool in an industry of exploiting and then discarding the talent that creates the material that makes them the money they don't share fairly with the creatives.

If there was a way to train and use AI without unfairly exploiting the labor of others, then we could begin to talk about the pros of using it. But we are not there yet, and because of the goals and inclinations o those promoting the use of AI in these situations, we likely never will be.

-M

[Mister Mets]

This is an important question, since we're talking about technological changes that may be as, or more profound as the shift from print to digital.

I'm not persuaded by arguments about taking jobs away. That's like calls to ban the lightbulb to support the candle industry.

The copyright questions are important, although that's probably going to be resolved. One solution would be to mandate that AI use material that is public domain, or that they've licensed. So in this case, if someone wants to draw using an Alex Maleev filter style, Alex Maleev is compensated.

There are some problems in how AI currently works. Much of it indiscriminately gets material from the internet, and that material may be copywritten. And they're designed to learn in a way where the programmers don't know what's going on. So it would be complicated to make it stick to particular parameters on copyright, but it may be necessary. Or they'll have to make a case that this is so fantastic it's worth the tradeoffs and maybe a system where artists whose work is used by AI get compensated by the taxpayers.

I've said this elsewhere but AI is a tool. I'd expect that many of the best artists are going to be able to use it well, just as they were more likely to make the best out of any new tool.

In comics, I expect AI will allow some artists to draw material faster. That will lower the cost of making comics, which means that products can be cheaper or appeal to a more niche audience. That's great for marginalized groups, when it becomes profitable for people ot tell their stories.

It's always going to take more time and require specialist knowledge to make something good, so we're not going to have a situation where one guy is pumping out a hundred comic books a month (or at least one hundred comics worth reading.)

[CaptainEurope]

My worry is, and this might sound arrogant to some:

Companies will totally put out comics with AI art. And the quality will be noticeably worse to those of us who notice these things and/or care out them. But there are enough philistines who will buy it anyway.

I mean, Greg Land has fans. And one time I was posting on social media about art that looked weird in a JSA mini and got attacked by fans saying I just wasn't getting the bold new style, until the actual artist came on and apologized for the reproduction error making the art look bad.

tl, dr: AI can't draw hands, Liefeld can't draw feet, but people buy all of it anyway.

[Spike-X]

I won't be buying comics that aren't made by human beings.

[Mister Mets]

For the hell of it, I decided to try an AI art generator. I went with Bing Creator, since it seems to be well-regarded and it's free.

The image quality varies a lot. A fan of Spider-Man, Sonic the Hedgehog and New York City, I asked for pictures of Spider-Man chasing Sonic the Hedgehog at Coney Island. Bing Creator approximates Spider-Man consistently, but Sonic not so much.



For this one, it kind of made a Spider-Man with Sonic's mouth. And Sonic is certainly not on-brand.



A very different thing I asked for was a World War 2 biplane flying over Medieval Tallinn (capital on Estonia) in a pop art style.



It's odd to figure out the language to talk about this. I keep being tempted to say that I made it, but I didn't. I asked for it.

It is one thing to get an image that makes sense by itself, or as part of a montage. This is still nowhere no close enough to the sequential storytelling at the heart of comics. But AI may very well get better.

[Mister Mets]

This was a nunnery with three people in a forest. It came out okay in one image, even if it got the number of nuns wrong.



This nunnery seems unholy.



These poor sisters appear deformed.

[ed2962]

I went to Bing and put in Texas Chainsaw Garfield the cat and got an image. I put in Batman kissing Superman they wouldn't do it. I put in Heavy Metal cheerleaders they wouldn't do it. I put in heavy metal bikers, I got an image. I put in Punk Rock Cheerleaders and got an image. This is one of the images I got when I put in Disco Dazzler and Wolverine...

[KurtW95]

As I said in the OP, I hear that the parts that screw up faces and other aspects are purposely put in there on free versions and those that everyone can use.

[Mister Mets]

As I said in the OP, I hear that the parts that screw up faces and other aspects are purposely put in there on free versions and those that everyone can use.

Whether free programs are nerfed is a bit of a moot point. Current problems will likely be improved in the future.

I'm not sure screwed up faces makes sense as part of what distinguishes the premium version. Doing that kind of thing without being clear about it would hurt the promotional value of the free versions.