D&D General Would It Matter To You if D&D Books Were Illustrated by AI Instead of Humans?

[Bluebell]

You know, one of the things that motivates me to purchase an expensive hardcover guide rather than simply looking up its contents online is the artwork. I see physical guides as sort of coffee table artbooks as much as they are guides. Why would I pay money for AI-generated artwork rather than clicking around an AI website myself? Why should I place any value on a product that devalues artists so completely as to cut them out?

 

[the Jester]

Would it ruin the art? No. But I'd much rather see humans getting paid to do the art on my new dharculus entry in the Monster Manual, or on the new pic of the tabaxi pc write up.

 

[EzekielRaiden]

Yes, it would bother me.

Artists deserve to be paid for their work.

It would not bother me because it is AI, that's whatever. Some AI-generated art is quite interesting, even beautiful. But unless we turn "art AI" into a tool used by artists and thus a tool which enhances artists' ability to create art, rather than a tool which enhances tech companies' ability to milk money out of us, it will be a Problem and not a Solution.

 

[Ruin Explorer]

It matters because it's essentially semi-legal theft and eventually, it'll be a matter of still needing actual human artists to create new styles, new approaches and so on. The AI can only replicate what already exists, and whilst it might get better at that, it's just emulation.

So personally I would not support any paid product using it, including small operations, because you're basically ensuring artists, who have a tenuous employment at best, have no employment at all.

As others have said, it doesn't bother me that it's AI. It bothers me that it's ensuring artists have even fewer employment prospects, and the exciting art is even less likely to be supported finanicially.

 

[Ruin Explorer]

Yeah, at this rate AI image generators will be able to outperform any human artist probably within the year, both in quality and speed.

Also no, no they won't, not in a year. In ten or twenty years? Maybe. Speed they'll obvious be ahead, but the "custom character portraits" I've seen using them so far have been pretty poor, and could in many cases be equalled or improved upon simply by going to Pinterest or er that big concept art site I'm forgetting the name of.

 

[EzekielRaiden]

AI images for photorealistic faces: thispersondoesnotexist.

Refresh the screen for a new face. Each face is unique. Make sure you save it if you want it, or it is gone.

Though, it's worth noting, it does secondary faces much, MUCH more poorly than primary faces. The face at the center of the frame will almost always look perfect. A face at the edge of frame will almost always look wrong.

And that's saying nothing about the surrealist backgrounds or clothing I've seen.

Yeah, at this rate AI image generators will be able to outperform any human artist probably within the year, both in quality and speed. I'm already seeing people use them to get custom character portraits of their D&D characters much faster and cheaper than they ever could have by commissioning a human artist.

The main issue with doing this is, you'd better hope you want stuff that's really common, even universal. "Old man wizard," for example, will be perfectly easy. If you like dragonborn, on the other hand? Good friggin' luck. AI-generated art of more overtly non-human figures is rarely any good. And if you want something more complex, like an actual dragon or a manticore or something? Don't even bother. Every AI I've used to try to get stuff like that spits out nothing but useless garbage.

 

[EzekielRaiden]

Also no, no they won't, not in a year. In ten or twenty years? Maybe. Speed they'll obvious be ahead, but the "custom character portraits" I've seen using them so far have been pretty poor, and could in many cases be equalled or improved upon simply by going to Pinterest or er that big concept art site I'm forgetting the name of.

Most likely ArtStation. And yeah, completely agreed.

 

[Ruin Explorer]

The main issue with doing this is, you'd better hope you want stuff that's really common, even universal.

This is exactly right and illustrates the major problem with AI-driven art. It can only operate by working from existing art which it has access to. This why some prompts work wildly better than others. And the more mainstream the prompt, the better the result. I kind of wonder if this'll eventually reactively shape the choices of designers and players, if it does get used heavily. I mean, will people not pick Aardlings because AI art won't have any clue what that is, and will be working off like three images, initially (and even when it's 300, it won't be much compared to most ideas)? I don't think they will straight away, but over time? Hmmm.

And yes ArtStation, that's the one. I often use it to nick stuff to show the players, but also just for inspiration, because whilst there's an awful lot of cheesecake and boredom (oh god so much cheesecake!), there's sometimes some really amazing stuff on there.

 

[Lakesidefantasy]

Artificial Intelligence will eventually replace the Game Master. I think 2023 is a safe prediction for when well see that, if it isn't already a thing now. Hasn't AIDungeon already been tasked with that?

This is all part of the Second Digitization.

 

[gnarlygninja]

It'd bother me quite a lot, it would almost definitely be done as a way to save money by not having to pay artists and I don't really care how much revenue each book generates for WotC. It's not like those savings would be passed along to us consumers anyway, and artists deserve to earn a living.