D&D 5E Glory of the Giants' AI-Enhanced Art

The latest D&D sourcebook, Bigby Presents: Glory of the Giants, comes out in a couple of weeks. However, those who pre-ordered it on D&D Beyond already have access, and many are speculating on the presence of possible AI art in the book.

One of the artists credited is Ilya Shkipin, who does traditional, digital, and AI art. In an interview with AI Art Weekly in December 2022, Shkipin talked at length about their AI art, including the workflow involved.

On Twitter, Shkipin talked more [edit--the tweet has since been deleted but the content is below] about the AI process used in Bigby, indicating that AI was used to enhance some of the art, showing an example of the work.

There is recent controversy on whether these illustrations I made were ai generated. AI was used in the process to generate certain details or polish and editing. To shine some light on the process I'm attaching earlier versions of the illustrations before ai had been applied to enhance details. As you can see a lot of painted elements were enhanced with ai rather than generated from ground up.

-Ilya Shkipin​

​

​

Discussions online look at more of the art in the book, speculating on the amount of AI involvement. There doesn't appear to be any evidence that any of the art is fully AI-generated.

AI art is controversial, with many TTRPG companies publicly stating that they will not use it. DriveThruRPG has recently added new policies regarding transparency around AI-generated content and a ban on 'standalone' AI art products, and Kickstarter has added similar transparency requirements, especially regarding disclosure of the data which is used to train the AI. Many artists have taken a strong stance against AI art, indicating that their art is being 'scraped' in order to produce the content.

UPDATE- Christian Hoffer reached out to WotC and received a response:

Have a statement from Wizards over the AI enhanced artwork in Glory of the Giants. To summarize, they were unaware of the use of AI until the story broke and the artwork was turned in over a year ago. They are updating their Artist guidelines in response to this.

Wizards makes things by humans for humans and that will be reflected in Artist Guidelines moving forward.

-Christian Hoffer​

​

The artist, Ilya Shkipin, has removed the initial tweet where the AI process is discussed, and has posted the following:

Deleted previous post as the future of today illustrations is being discussed.

Illustrations are going to be reworked.

-Ilya Shkipin​

​

 

[Uni-the-Unicorn!]

I'm quite aware of how quick it's going. Trust me, I'm following it fairly closely. But at the same time, with great strides come setbacks, too. We were mocking AI for its problems with hands (and teeth, which was another big tell), but as it has gotten better at that it still has problems, even when just "enhancing" things. I would definitely agree that using the AI first and then having the human touch is probably the better workaround, but the inherent problem there is that getting the exact product you want is kind of a crapshoot with AI.

It seems to me that Ilya’s approach is actually better both in results and ethics to me - if I understand correctly.

Step 1: artist makes a draft
Step 2: AI used to “enhance” draft
Step 3: artist finishes draft

You can repeat steps 2-3 as needed. Personally, I much prefer AI used to enhance than create the concept from the beginning. Heck, I’m good enough to use photoshop and “correct” the artifacts in a lot of AI generated art. But I definitely don’t have the talent to make any of Ilya’s original sketches.

 

[Uni-the-Unicorn!]

It could. But it doesn’t. If it did, there would be no ethical issues at all.

That it can is important though. Also, are you sure it is not being used that way? If not here, but by someone somewhere? I know of people who have used AI to make written work in their “style” be feeding it their own work. Seems like something you could do with art too.

 

[Justice and Rule]

It seems to me that Ilya’s approach is actually better both in results and ethics to me - if I understand correctly.

Step 1: artist makes a draft
Step 2: AI used to “enhance” draft
Step 3: artist finishes draft

You can repeat steps 2-3 as needed. Personally, I much prefer AI used to enhance than create the concept from the beginning. Heck, I’m good enough to use photoshop and “correct” the artifacts in a lot of AI generated art. But I definitely don’t have the talent to make any of Ilya’s original sketches.

I mean, it's better than just making it up wholesale, but better doesn't mean "good".

The problem is Step 2: to "enhance", the AI needs a dataset to go off of. Unless Ilya has a rather extensive database of art that he has created that works within the style, he's probably getting it from somewhere else. That's where the problem comes in.

 

[Golroc]

Have you seen the concept art and the final works? Someone posted them a few days ago and they final works do not appear to be copies of the concept art. They are original works IMO.

No, I haven't. In that case, that's a good thing. Because it would be really odd to just transform concept art into an illustration. That's not really how concept normally works (at least not in computer game development).

 

[Morrus]

That it can is important though.

Yes. It is. That’s why I’ve been suggesting ethical opt-in data sources, disclosure of data sources, and regulation for quite a while now!

Also, are you sure it is not being used that way? If not here, but by someone somewhere? I know of people who have used AI to make written work in their “style” be feeding it their own work. Seems like something you could do with art too.

I’m sure someone somewhere is doing anything you can think of.

 

[Uni-the-Unicorn!]

I mean, it's better than just making it up wholesale, but better doesn't mean "good".

The problem is Step 2: to "enhance", the AI needs a dataset to go off of. Unless Ilya has a rather extensive database of art that he has created that works within the style, he's probably getting it from somewhere else. That's where the problem comes in.

Sure and it also, in my mind, depends on what the enhancement is. I’m still interested in the idea of an artist using AI that only samples their own art. In think WAR could pull that off

 

[Jfdlsjfd]

Sure and it also, in my mind, depends on what the enhancement is. I’m still interested in the idea of an artist using AI that only samples their own art. In think WAR could pull that off

You'd need a much larger database. What is doable is use a base model and create a derivation base on his own works -- much like the same process where people take pictures of them in several positions and clothes and making faces to train the IA to generate image of them on the moon.

 

[Uni-the-Unicorn!]

You'd need a much larger database. What is doable is use a base model and create a derivation base on his own works -- much like the same process where people take pictures of them in several positions and clothes and making faces to train the IA to generate image of them on the moon.

That is basically what I’m talking about if I am understanding you

 

[Jfdlsjfd]

That is basically what I’m talking about if I am understanding you

Then yes, it's possible, and can be done with a few images. 20-50 to train the concept of an item or person is usually enough, it would be more (but still reasonable) to train a style. The work would be in captioning the images (so the model would understand that you're asking to orient generations toward what is your "style" and not toward, say, drawing cats because most of your images were cat pictures). A good captioning will help a lot, and it could be done by hand by the artists if they wanted to.

Not sure if Adobe will go this way with Firefly, allowing users to train their own style to have an ethical AI (whatever that means) that generates as they'd draw, but it could be a selling point for some. Maybe too few to warrant going this way, though.

 

[Justice and Rule]

Sure and it also, in my mind, depends on what the enhancement is. I’m still interested in the idea of an artist using AI that only samples their own art. In think WAR could pull that off

I mean, in concept you could do it. It'd be ethically better but I feel like I'd be kind of ambivalent on it? But at the least, it would be a "do no harm" sort of scenario. But you'd need a large dataset to do that.