WotC WotC: 'Artists Must Refrain From Using AI Art Generation'

After it was revealed this week that one of the artists for Bigby Presents: Glory of the Giants used artificial intelligence as part of their process when creating some of the book's images, Wizards of the Coast has made a short statement via the D&D Beyond Twitter (X?) account.

The statement is in image format, so I've transcribed it below.

Today we became aware that an artist used AI to create artwork for the upcoming book, Bigby Presents: Glory of the Giants. We have worked with this artist since 2014 and he's put years of work into book we all love. While we weren't aware of the artist's choice to use AI in the creation process for these commissioned pieces, we have discussed with him, and he will not use AI for Wizards' work moving forward. We are revising our process and updating our artist guidelines to make clear that artists must refrain from using AI art generation as part of their art creation process for developing D&D art.

-Wizards of the Coast​

Ilya Shkipin, the artist in question, talked about AI's part in his process during the week, but has since deleted those posts.

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 Shlipin​

​

 

[Art Waring]

Because it's not a requirement. It's just a happenstance that Stability had a copyright-free stock of music (old music records)? Also, generating randomly a few seconds of recognizable existing music is certainly more statistically possible than generating a existing image. If I sing a few seconds of a music of an easily recognized score (one can recognize Twinkle Twinkle Little Star with 14 notes easily, and the odd of generating it are higher than a graphic generator generating a recognizable part of the Mona Lisa, barring overtraining on Mona Lisas.

So you are saying that since its not currently a requirement, that this is set in stone? The laws can change, and with them the guidelines for how to create ethically sourced datasets in the future.

All I'm saying is they could have gone about this a bit better than saying "its not mandatory," hopefully avoiding this whole mess entirely.

 

[Sorcerers Apprentice]

So you are saying that since its not currently a requirement, that this is set in stone? The laws can change, and with them the guidelines for how to create ethically sourced datasets in the future.

All I'm saying is they could have gone about this a bit better than saying "its not mandatory," hopefully avoiding this whole mess entirely.

I'm not sure if it's legally possible to restrict what such a dataset is used for. It's essentially just a list of image URLs and tags, that's not a creative work, and as such it's not eligible for copyright protection.

 

[Jfdlsjfd]

So you are saying that since its not currently a requirement, that this is set in stone? The laws can change, and with them the guidelines for how to create ethically sourced datasets in the future.

No I am not saying that the current situation is optimal. I am discussing what is, currently, the state of what is possible or not. Of course laws can change, and with 190+ countries on Earth, there is a strong chance all the spectrum of outcome will exist. Laws changed when copyright was created (initially at the benefit of printers, not authors, in Ireland, for example), when copyrighted shifted to benefit the author, when fair use was created as an exception to a strict monopoly, when the moral rights (indefeasible) were separated from the commercial rights (with a statute of limitation), when said statute was extended, when WWI happened and it became necessary to extend the benefits of copyright because of the disruption of the wars, when it was decided that database can be copyrighted as work of the mind and so one. However, there is no telling how it will evolve. Will it include training as fair use? Will it ban AI altogether? Will it consider the need of Adobe and Google over the need of the individual artists? Will it consider the benefits of having free, average quality, art for everyone and make it a public service free for all like education? With so many countries, it's already diffcult to assess what is legal right now, and IP laws are rather converging, without speculating over what will happen and where.

Also, the laws wouldn't necessarily align with what is ethical, even if could achieve a consensus about what is ethical, but that's a topic that's necessarily political and forbidden here, so I won't go into this.

 

[Jfdlsjfd]

I'm not sure if it's legally possible to restrict what such a dataset is used for. It's essentially just a list of image URLs and tags, that's not a creative work, and as such it's not eligible for copyright protection.

That would really depend on where you are. There is a separate provisions for copyright of databases in the US where, say, a database of quotes can be protected because there is a creative work in the selection of quotes. Whereas in Europe, it would be protected irrespective of the creativity but it would require being new and necessitate substantial effort, and the protection would only last for 15 years.

In the current state of things, they claim it's an uncurated list of items, so it doesn't display a lot of creativity. However the description of the images may be original (?).

 

[FrogReaver]

Actually, that is an incredibly surprising result, despite your post hoc rationalization.

Look, if the AI had centered on, say, "Ivy league degrees," we might say that this is unfair, but understandable.

In that context it is more surprising.

But if you went back to the people who actually made the decisions that the data set was based on, and told them- "Hey, I know you think you're being fair. But do you know what the two most important factors are?

I’m not going to keep adding speculation - but there’s a lot of possibilities for why it came out that way and without some assurances that those other possibilities were cared for then I think we are kind of at square one. People are less critical of anything that aligns to their already held views.

Whether the applicant is named Jared, and whether they were a Lax Brah ...." I think that they would probably be more than a little surprised. Again, this isn't even college lacrosse ... these are people that listed high school lacrosse on their resumes.

Or Bill/Richard. Or Greg. I think all were fairly common names at one point but didn’t show up as the name above all others this time.

If it’s all about a specific name, why is it that the particular name changes as time marches on.

In other words, it did uncover something rather surprising. Which should lead to some self-reflection. Whether you think it should lead to self-reflection or not ... well, I can't control what you choose to do, right?

Lol. I don’t find it surprising.

 

[Jfdlsjfd]

I'm not sure if it's legally possible to restrict what such a dataset is used for. It's essentially just a list of image URLs and tags, that's not a creative work, and as such it's not eligible for copyright protection.

Oh, and I checked about LAION-5B over the Internet, it was created by a German high school teacher as a pet project over less than a year. While he probably didn't count the cost of his own labor,

The largest dataset for training text-to-image generators was assembled by volunteers for roughly $10,000.

Source: The Story of LAION, the Dataset Behind Stable Diffusion

Honestly, for this sum of money, a lot of private persons could do the same, and any group interested to have access to the same database could invest it easily, in a world where kickstarters make millions. I am pretty sure it could be raised easily by a group promising nsfw waifus generator or "input your character background, get 300 images to choose from" users.

So Stability would have just replicated the work in house if they could be deprived of the benefit of LAION for their activity.

 

[Golroc]

It’s amazing how you can be so right and so wrong at the same time.


Close enough. They don’t understand any relationships. They simply have identified that relationships exist. To me that’s a big difference.

Yet that’s exactly where their training data came from.

It's amazing you can be so passive-aggressive and insulting against someone who tries to contribute value and information to a thread that is full of misinformation and ignorance. But thanks for driving me off from this forum. If this thread is representative of these forums, I'd better leave before I waste any more time on such a toxic community. You should be ashamed of yourself.

Edit: Got a bit riled up there. Looks like my impression was wrong and people are perhaps more level-headed. Being able to give an apology is a quality I appreciate (not being sarcastic), so I'll stick around and see what else this place has to offer.

 

[Umbran]

It’s amazing how you can be so right and so wrong at the same time.

Mod Note:
Dude. Can you roll the disdain back several notches, please? You are talking to actual people, not AIs. The people matter.

 

[FrogReaver]

It's amazing you can be so passive-aggressive and insulting against someone who tries to contribute value and information to a thread that is full of misinformation and ignorance. But thanks for driving me off from this forum. If this thread is representative of these forums, I'd better leave before I waste any more time on such a toxic community. You should be ashamed of yourself.

Apologies. I was.

 

[nevin]

How does a machine define "style", except by reference to examples?

Even if the operation were like you say, the training of the machine requires making digital copies of the works, which is generally a copyright violation right there.

Thank the techbros, corpos, and media for pushing this stuff with as much sweet, sweet sensationalism as the We the Consumers can handle!

the easiest way to get away with something is get everyone focused on something else.