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 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.

I am curious how large the dataset would need to be. My son is in this field so I will have to get his thoughts.

 

[Jfdlsjfd]

I am curious how large the dataset would need to be. My son is in this field so I will have to get his thoughts.

It depends. The more accurate the captioning, the less images you'll need, generally. Also, the more image, the more of your style will be identified. But only you will be able to judge if an image is in your style or not. The usual ways to go is to generate X images and select one that is the closest to your goal before working on it. If you can "afford" generating 64 images to choose from instead of 8, you can probably use less images. People who tried claim to be happy with 200-400 images, but it depends on expectation.

 

[tomBitonti]

AI is really good at looking impressive at a glance. Once you look closer, it's like realizing you're looking at an illusion: you begin to see the artifice and the rest of it starts to fall apart. Same deal with the pictures here: they look okay on a glance, but if you actually look at what's going on, it starts to fall apart.

Edit: I think the best way to put it is that AI is good enough at getting close to what we think should be there that our minds can fill in what we think is there and think it's really good... until we look and see that it's not what we were mentally filling in, but something far less refined and far more chaotic.

Although, isn't that the case with a lot of art? In particular, painters seems to do a lot of providing just enough form and just enough detail and relying on our perception and our brains to fill in the rest.

TomB

 

[Justice and Rule]

Although, isn't that the case with a lot of art? In particular, painters seems to do a lot of providing just enough form and just enough detail and relying on our perception and our brains to fill in the rest.

TomB

Not really in the same way or to the same extent. Just like the conversation I had regarding mistakes, the mistakes AI make are very different than the shortcomings of an actual artist.

 

[Jfdlsjfd]

Although, isn't that the case with a lot of art? In particular, painters seems to do a lot of providing just enough form and just enough detail and relying on our perception and our brains to fill in the rest.

TomB

Those pointilist did suck at drawing details, I can tell you!

(Fun anecdote: I was once at the free museum night and overheard a couple looking at a 19th century realistic painted, and a person said to the other: they were much more precise than in the other (pointilist) room, art had really advanced a lot!).

 

[tomBitonti]

Not really in the same way or to the same extent. Just like the conversation I had regarding mistakes, the mistakes AI make are very different than the shortcomings of an actual artist.

Using the example of the car spokes, is it really a mistake? The image seems to work even with incorrect spoke details. I’m thinking technical inaccuracy doesn’t matter as much as whether our perceptions accept an image; or whether an image is evocative. We are not counting ships in an aerial reconnaissance image.
TomB

 

[Vaalingrade]

I mean, really is a thing that has not actual intentions and is only following its orders even capable of making something we can consider a 'mistake'. A human artist knows what the spokes are meant to look like. The computer only has its programming. It doesn't know anything.

 

[Justice and Rule]

Using the example of the car spokes, is it really a mistake? The image seems to work even with incorrect spoke details. I’m thinking technical inaccuracy doesn’t matter as much as whether our perceptions accept an image; or whether an image is evocative. We are not counting ships in an aerial reconnaissance image.
TomB

Like @Vaalingrade says, humans can make things inaccurate out of intention or style or simply not knowing. This isn't meant to be evocative, this is just a program putting pieces together and making artifacts because it doesn't realize that car spokes have to be straight or railings don't look like frayed threads. Yes, we are not counting ships in recon photos, but also human artists wouldn't put nonsense shapes like in the driver's seat of a car like we see in that image. They aren't comparable.

 

[Jfdlsjfd]

I think "mistake" can be replaced by "problem" and @Timbitonti observations is still meaningful. Sure, Disney drew Mickey with 4 fingers and it's not because Walt didn't know that people (and actual mice) have 5 fingers, or because he doesn't understand the concept of fingers and its interactions with the concept of hand. He was able to. He just chose to draw four because it was easier to animate and because, ultimately, it doesn't matter to the audience. AI does thing wrong, but as long as it doesn't matter, we don't care. In films, wheel spokes go in the wrong direction more often than not, and we simply ignore it and can process the idea that it represents a car moving along a road, even if it should logically go in the other direction, by a abstracting and taking into account all the other elements from the scene. And if more than a few people aren't distracted from enjoying films despite their obvious wheel spoke inadequation, more than a few people will tolerate uneven railing, especially for something as trivial as drawing an illustration of our make-believe-elf-wizard (If I've understood Tim's intention correctly).

 

[pukunui]

@Morrus: There’s a new community update on D&D Beyond addressing this issue. In it, WotC states they have removed all the Bigby art involving generative AI from DDB and have commissioned replacements. These will appear on DDB soon and in future printings of the physical book (so anyone who gets a first printing will still have the AI ‘enhanced’ art).