RPG Evolution: The AI DM in Action

How might WOTC launch an AI-powered DM assistant?

Picture courtesy of Pixabay.​

We know Wizards of the Coast is tinkering with Artificial Intelligence (AI)-powered tools for its multiple properties, including Dungeons & Dragons. But what might that look like in practice?

Interactive NPCs​

Large Language Model (LLM) AIs have been used extensively to create non-player characters of all stripes on Character.AI. It's not a stretch to imagine that Wizards might have official NPCs included as part of the digital purchase of an adventure, with the rough outline of the NPC acting as parameters for how it would interact. DMs might be able to create their own or modify existing NPCs so that the character drops hints or communicates in a certain way. Log outputs could then be available for DMs to use later.

There are several places today where you can create NPC bots powered by AI that are publicly available, although the DM might need to monitor the output in real time to record the conversation. Character.AI and Poe.com both provide the ability to create publicly available characters that players can interact with .

Random Generators​

There are already dozens of these in existence. What's particularly of note is that AI can go deep -- not just randomize what book is in a library, but provide snippets of text of what's in that book. Not just detail the name of a forgotten magic item, but provide stats for the item. For WOTC products, this could easily cover details that no print product can possibly encompass in detail, or with parameters (for example, only a library with books on necromancy).

AI RPG companion is a great example of this, but there are many more.

Tabletop Assistants​

Hasbro recently partnered with Xplored, with the goal of developing a "new tabletop platform that integrates digital and physical play." Of particular note is how Xplore's technology works: its system "intelligently resolves rules and character behaviors, and provides innovative gameplay, new scenarios and ever-changing storytelling events. The technology allows players to learn by playing with no rulebook needed, save games to resume later, enables remote gameplay, and offers features like immersive contextual sound and connected dice."

If that sounds like it could be used to enhance an in-person Dungeons & Dragons game, Xplored is already on that path with Teburu, a digital board game platform that uses "smart-sensing technology, AI, and dynamic multimedia." Xplored's AI platform could keep track of miniatures on a table, dice rolls, and even the status of your character sheet, all managed invisibly and remotely by an AI behind the scenes and communicating with the (human) DM.

Dungeon Master​

And then there's the most challenging aspect of play that WOTC struggles with to this day: having enough Dungeon Masters to support a group. Wizards could exclusively license these automated DMs, who would have all the materials necessary to run a game. Some adventures would be easier for an AI DM to run than others -- straightforward dungeon crawls necessarily limit player agency and ensure the AI can run it within parameters, while a social setting could easily confuse it.

Developers are already pushing this model with various levels of success. For an example, see AI Realm.

What's Next?​

If Hasbro's current CEO and former WOTC CEO Chris Cocks is serious about AI, this is just a hint at what's possible. If the past battles over virtual tabletops are any indication, WOTC will likely take a twofold approach: ensure it's AI is well-versed in how it engages with adventures, and defend its branded properties against rival AI platforms that do the same thing. As Cocks pointed out in a recent interview, WOTC's advantage isn't in the technology itself but in its licenses, and it will likely all have a home on D&D Beyond. Get ready!

 

[Oofta]

Legally speaking, yes, it is intrinsically different. I'll get to that in a second.



Sure. It isn't like human artists' work never infringe - the whole idea of infringement predates computers by centuries! The line between infringing product, and not, is a fuzzy one, and you have to start making weird "points of similarity" arguments to decide which side of the line a work is on.

However, we may not need this to make a case against current generative AI.



You're looking at the wrong end of the animal. Don't look at the output. Look at the input.

The first (and possibly best) legal argument here has nothing to do with the details of the end product. And the music industry, in trying to kill Napster, did the heavy lifting for us already. And it is this: making an unauthorized digital copy of an artwork is considered copyright infringement.

And, long before the generative AI makes any product, even before it is trained, the folks trying to train it must make digital copies of the artworks to include them in the training set.

This is specifically why most of the early generative AIs were presented to the public "for academic use" - because academic use is usually covered by Fair Use doctrine, so that the training copy could be overlooked.

It is this act - scraping the internet and making digital copies of works as parts of the training set, that is a problem.

Again, just pay the darned artists, and this goes away.

Comparing Napster to what AI does is kind of comparing apples and oranges. Napster took a song they didn't pay for without modification and allowed people to listen to it. Art (and text) is used to build up pattern recognition and associations. Much like a person does reading books, studying existing art or for that matter, interacting with the world around them. AI couldn't do what it does without input, but it is not simply copying art. Ideally you could compensate the original artist somehow but that would be an impossible task, the myriad images come from all over and there's not really a way to trace it back.

LLMs don't think like humans do of course, and even if we ever get a true AGI it likely won't think like a human either. But it reminds me of the argument that people are somehow unique and special animals that we "think" different that any other animal on the planet. Which the more we learn about animal cognition, the more we realize it's not true. It may be a matter of scale and scope but our brains are not fundamentally different. We all just take in input, process it to learn to recognize patterns and structures. AI does the same, just in a far more limited scope and using different methods. It doesn't think, doesn't understand what the images represent to us, but we only recognize what the image is and give it meaning because of our own built in pattern recognition processes.

 

[Thomas Shey]

This is a false equivalency that rests on the mistaken assumption that human brains work the same as AIs or LLM.

No. Its an equivalency that doesn't consider the distinction important. At least take shots at the right thing.

 

[Umbran]

I'd be willing to put a side bet that you could pay the non-public sources for a copy of each of their works per each training process, and there'd still be strong pushback, because you see all too much of the objections having only a little to do with the scraping.

We get "strong pushback" on the idea that the world is round. If you are not going to be satisfied unless "strong pushback" is eliminated, you're going to be disappointed.

Again, nobody asks how legal the copies of various music or art human artists use when teaching themselves to do their art, whether they end up selling their work or not.

They don't have to ask, because Fair Use covers educational use. It is a special case that is very specifically and explicitly carved out in copyright law.

"...the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright." - 17 U.S.C. § 107

Were you not aware of this?

In an art class, a teacher can find a frame from Disney's Pocahontas online, make a big poster of it to put up in class, and ask all their students to reproduce it in oils as a practice in color use, and that's perfectly legal - both the poster and the student's copies - and the House of Mouse couldn't touch them.

Selling those paintings would make it no longer be Fair Use.

 

[Thomas Shey]

We get "strong pushback" on the idea that the world is round. If you are not going to be satisfied unless "strong pushback" is eliminated, you're going to be disappointed.

When I expect it would make no visible difference in the arguments seen, afraid I at least expect more.

They don't have to ask, because Fair Use covers educational use. It is a special case that is very specifically and explicitly carved out in copyright law.

"...the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright." - 17 U.S.C. § 107

Were you not aware of this?

In an art class, a teacher can find a frame from Disney's Pocahontas online, make a big poster of it to put up in class, and ask all their students to reproduce it in oils as a practice in color use, and that's perfectly legal - both the poster and the student's copies - and the House of Mouse couldn't touch them.

Selling those paintings would make it no longer be Fair Use.

And as I've made clear, I don't consider that meaningfully different than what an AI is doing with the same material. It probably is legally because of how its fed to the AI, but I'm remarkably uninterested in what the law says on these sort of discussions.

As I've said "Duplicate X piece or art by Person Y" should be illegal in either case. "Look at this particular artist and imitate his style" is quesitonable in either case, but happens all the time. "Be influenced by his style in a way that shows up in your work" is not, and should not be, and I don't think there's any meaningful difference whether that's done by a neural net or a human. I'm aware others disagree with me, as is their right, but I'm not required to change my position because they do.

 

[Umbran]

Comparing Napster to what AI does is kind of comparing apples and oranges. Napster took a song they didn't pay for without modification and allowed people to listen to it. Art (and text) is used to build up pattern recognition and associations.

So long as the Generative AI was being used for educational and research purposes only, they are apples and oranges. Once you're trying to make a buck off it, however, the differences between them are not particularly important - they are now both oranges.

 

[Thomas Shey]

So long as the Generative AI was being used for educational and research purposes only, they are apples and oranges. Once you're trying to make a buck off it, however, the differences between them are not particularly important - they are now both oranges.

I see the distinction you're making, but again, if you hire an artist or musician, nobody asks what works they used when they were learning to do their work.

 

[Oofta]

Yes.

You can get world class art for $60. (I've commissioned absolutely tons of D&D art. I just spent about $2200 on a commission).

I will never pay $60 for a picture for D&D. I've paid far more for miniatures that I've painted over the years, but to be honest even that process is changing. I used to buy metal minis, now if it's a PC we use heroforge, for monsters I download them from someone* who does the stls and I print them at home. Do I feel a bit bad for no longer supporting my local gaming store and reaper minis? Yeah, but for me the end result matters more. Times change. I still paint the minis myself and they mean more to me because I've painted them, but the reward of painting is in the process and result itself.

When it comes to pictures on my wall, most are photos I've taken while on vacation. If I ever do start making images for my group, it will be using AI as a tool and the results would mean more to me because I had input into the process.

*Get more from mz4250 on Patreon

 

[ART!]

Noodling around with an AI image generator until you get exactly what you want is arguably not making art, because a) art isn't about getting what you want, and b) no artists winds up where they wanted when they first started a piece.

 

[Raiztt]

Yeah, but for me the end result matters more.

Consumerism in a nutshell A real meme/comic moment in the wild

 

[Oofta]

Consumerism in a nutshell

Yep, I am a consumer. If I'm getting something I would otherwise not purchase I don't see an issue.