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

[EzekielRaiden]

Not the way they worded it. Including things like Inkcarnate? Their phrasing implies that anything digitally assisted needs to have that disclaimer, including anything made in photoshop, InDesign, ProCreate, and others. Which is ironic, because they themselves push templates for you to use, using those tools.

They are going to end up 95% of their products with that disclaimer. Not very well thought out at all.

I meant the underlying idea. Had not considered the implications there, so yes, the implementation is bad, but the notion of asking creators to specify in advance that they used AI art generation is good.

 

[GMforPowergamers]

The couple of times I tried those things with human figures I got some serious body horror, but I'm sure they'll fix it eventually.

I mean maybe that would be cool for a aberration book

 

[EzekielRaiden]

You folks are talking as if AI results are magic, and fall out of the aether for free.

There are a bunch of very talented people behind those AIs - I would expect they need to be very talented at code, as well as at art. And they've had to work very hard to get the results they do.

A person who is talented at art, using their skills to produce art... is an artist, no?

Sure. But the rewards of the work will go to the tech company, generally speaking. Unless you mean to say that journalists and authors are a large portion of people working inside Google and other AI platform places....which I highly highly doubt. Part of the problem is that you don't need to be an artist to train an AI on other people's art.

Edit: and before anyone gets at me about how their art is based on other art - I will direct you at the idea of copyright. Copyright has a limit so other folks can eventually base other art upon a given work. Licensing is similar. So, we are consuming art based on other art all the time.

Again the issue is not the provenance. I am in fact actually very much in favor of reducing copyright term limits. My issue is that these tools are almost guaranteed to get used by greedy corporations that want to cut costs wherever they can. If you can generate halfway decent art using DALL-E at one-tenth the cost (and time) of hiring an actual living artist, you do it, hands down, no question. Because the bottom line is everything, as we have seen from numerous problems arising over the past 30-50 years.

Automation can be a lovely thing. I made a comparison between a proposed "flat colors" tool and a washing machine earlier and I very much meant it. Washing machines have saved millions, perhaps billions, of people thousands of hours of labor over the course of their lives. Such tools are a good thing.

My concern is greedy corporations abandoning an industry that they already treat extremely poorly and exploitatively, chasing lower costs no matter the human cost, and as a result giving us more limited, samey, cookie-cutter art liable to be plagued by racist tropes and presentation because AI just use what we give them and there are major racist biases in both the data provided and the ways that data is used. (See: the image sharpener AI that turned pictures of black people into white ones. Including Barack Obama.)

 

[nyvinter]

There are a bunch of very talented people behind those AIs - I would expect they need to be very talented at code, as well as at art.

That do you base that on? What I've seen of the AI art companies are that they're started by people who want to be good at art but quit on day two because no noticeable skill change*. The machine learning software analyses what's popular in the image set, but it has no compositional weighting that I'd expect someone who was knowledgeable in art would put in.

*) Just look at all the rallying cries of "finally freeing art from those pesky artists!"

 

[Stormonu]

Somewhat...

I watched the new Beavis & Butthead movie. Despite the artwork being "better" (in that it was made using modern methods,) it took something away from the feel of the show.

This isn't limited to AI, but I find that a lot of contemporary art looks too clean. It's lacking something.

I don't know how to explain it, but I find less-perfect artwork more enjoyable.

That doesn't mean that I want bad artwork. But there's something about computerized art (and movie CGI) which doesn't look quite right. I don't know how to explain why or what seems off about it, but it's something.

I think what you are referring to is the "uncanny valley" effect. The artwork too closely mirrors the object that it reflects that it sits in an uncomfortable space between realistically mirroring the subject matter but being "off" at the same time.

 

[Sacrosanct]

AI art is a tricky thing. It's a conversation I see all the time in my woodworking groups re: tools. A lot of old timers get salty when someone makes something using a CNC machine as that's not "real" woodworking, only to be met with, "Unless you make everything with a rock, you're using tools as well. We all use tools, just different ones."

I think things like MidJourney are like that. I'm a bit (stressing a bit) of an artist myself, and most artists I know use tools like Procreate, and in those tools, we use things like blending effects, layers, etc. I'm not taking physical paint and mixing it to get a blend effect like I do when I paint minis--I'm letting the software do that work for me.

So part of me thinks AI art is to art world as CNC is to woodworking.

 

[MoonSong]

There are a bunch of very talented people behind those AIs - I would expect they need to be very talented at code, as well as at art. And they've had to work very hard to get the results they do.

No, Neural networks are basically pure math constructs. People creating a neural network bot don't really need to know -or bother- with the specifics of the thing the bot will work with. It isn't trivial per se, it takes a lot of code monking and creativity to make good bots, but the bot itself is a huge group of relatively simple units, and the actual final products depend solely on the training the bot has and the computer power that can be given to the bot. Source art major and Math/CS dropout.

 

[MoonSong]

AI art is a tricky thing. It's a conversation I see all the time in my woodworking groups re: tools. A lot of old timers get salty when someone makes something using a CNC machine as that's not "real" woodworking, only to be met with, "Unless you make everything with a rock, you're using tools as well. We all use tools, just different ones."

The CNC is still working using a design that a human created. A person had to take the decisions. AI art is instead replacing the human, not that there aren't humans giving maintenance to the bot, but they aren't making any decisions on the products of the bot. (However, I think it might be possible to have a bot that produces 3d models that in turn can be CNC'd/3d printed)

 

[Sacrosanct]

The CNC is still working using a design that a human created. A person had to take the decisions. AI art is instead replacing the human, not that there aren't humans giving maintenance to the bot, but they aren't making any decisions on the products of the bot. (However, I think it might be possible to have a bot that produces 3d models that in turn can be CNC'd/3d printed)

No, if you want a result that is anything other than atrocious, you have to input certain commands and parameters into AI art, just like you do with a CNC machine. You have to know at least a little about what you're doing to use MidJourney or other AI programs effectively, just like any other tool

It's the difference between this:

And this:

 

[doctorbadwolf]

AI art is a tricky thing. It's a conversation I see all the time in my woodworking groups re: tools. A lot of old timers get salty when someone makes something using a CNC machine as that's not "real" woodworking, only to be met with, "Unless you make everything with a rock, you're using tools as well. We all use tools, just different ones."

I think things like MidJourney are like that. I'm a bit (stressing a bit) of an artist myself, and most artists I know use tools like Procreate, and in those tools, we use things like blending effects, layers, etc. I'm not taking physical paint and mixing it to get a blend effect like I do when I paint minis--I'm letting the software do that work for me.

So part of me thinks AI art is to art world as CNC is to woodworking.

Absolutely this.

Imagine if you could just talk to Procreate and it would interpret your requested effect, and then ask if the result is what you intended.

It’s a tool. A very cool one, that lowers the barrier to entry and potentially expands what is possible to create. Imagine an artist create a work where the “AI” process of iteration and adding the resulting works back into the algorythm over to time, is part of the process.

Imagine an advanced art assistant that you can set up to record something, and “knows” how to link associated concepts like emotions and colors, feeling of movement from certain types of shapes, staccato rhythm represented visually, etc, and then translate a conversation into a piece of visual art.

These are all potential works of art that are enabled by this technology.

Now expand the scope. You are paying to have a house built, and your architect asks you a series of questions, and rapid fire shows you rooms and hallways and layouts, 2d and 3D, in hyper-detail, eventually constructing your exact dream house. The process still requires an architect with judgement who values human life and happiness, and it might always require at least review by a person, but what is practically plausible expands with the addition of this technology.