How AI Doesn’t Ruin TTRPGs

Despite having by far the most popular TTRPG product, Wizards of the Coast (WotC) has managed to consistently annoy their core audience over the past two years. The latest bit is a quote from Chris Cocks, the CEO of Hasbro (and previously President of WotC), who claims,

Inside of development, we’ve already been using AI… We will deploy it significantly and liberally internally as both a knowledge worker aid and as a development aid.

I’m not saying that social media is representative of community concerns, but the loudest voices do seem quite negative about the direction of D&D.

I don’t really know what’s going on over at WotC, but I’m optimist and hope that this isn’t really so bad.

So what’s the issue with AI?

Before I dive into this, let’s be clear what we’re talking about. AI is a broadly-used, constantly-changing term. Most of the talking points are about the use of Large Language Models (LLMs like GPT-4) and Generative AI (GenAI like ChatGPT or Midjourney).

The most obvious applications of GenAI to TTRPGs is to write adventures (or other content) and create digital artwork.

Let’s break down some of the concerns voiced by the community.

GenAI creates poor results

The images look bad. The adventures don’t make sense.

This very well may be true. I would hope that any final product from WotC, AI-generated or not, goes through significant editing, quality checks, play testing, and other review.

That being said, the community has criticized WotC products in the past when they didn’t use GenAI. The real question is whether WotC produces quality content, and we can only wait and evaluate each product as it is released.

DMs will lose their jobs

The specific quote from Cocks:

There’s not a single person who doesn’t use AI somehow for either campaign development or character development or story ideas. That’s a clear signal that we need to be embracing it.

Obviously he’s exaggerating, but his point is largely correct: GenAI is extremely popular with DMs. All DMs hope for that magic when, in the compost heap of the mind, they unleash true creative genius that changes players’ lives.

Or maybe you realize that tomorrow is game night, you have procrastinated on prep, and your players will have fun anyways.

However, the clear line is that GenAI is an aid, not a replacement, for DMs. I’m a software engineer, and when Copilot first came out, we wondered if the robots were coming for our jobs.

It hasn’t happened yet, and maybe it still will happen, but in the meantime, I write better code by using Copilot. It saves me writing out repetitive code. I learn things from how it works. I use it for exactly as much as I want.

And the same thing can happen in TTRPGs. As long as players want humans as their DMs, then there’s still demand for us. Even if AI can run better games than humans, we at least have an advantage in an innate desire to connect with other humans.

Artists and writers will lose their jobs (or souls)

Out of the criticism, I think this one has the most credence. In fact, they have already been caught on this point.

Perhaps more so than in other industries, tabletop gamers have a deep affinity for the craft. Simply by playing, we are engaging in the creative pursuit to make games come to life at our tables. When we see the art and stories in books, we want to believe that we have done good by supporting other, real humans in putting their passion into these products.

So when we see quotes like this that hint at the presence of GenAI within TTRPGs, our gut reaction is to see it as a threat to the creative spark and the seemingly fragile industry borne out of that.

Again, I think we can wait and see what happens with future WotC products. Even if we can’t tell by the final product, the community is small enough that we will know which artists stand behind their work or who was under pressure for it.

So why should we be optimistic about AI?

As I mentioned earlier, AI is a broad term, and even LLMs and GenAI have broad applications. Here are a few ideas I have on where this could go.

The art searcher

The internet has tons of wonderful art created lovingly by human artists that would be perfect for TTRPG use: portraits, tokens, battle maps, and more. It works, but I see two key problems.

First, attribution and payment. Once an image is out there, even a watermark can’t save someone from using it without payment. I don’t have a good AI solution for this, but just want to acknowledge it. Perhaps a centralized product streamlines and standardizes this process.

Second, images are really hard to search. Google image search is about the best general-purpose tool we have, but it certainly can’t take the equivalent of a Midjourney prompt and find real human results.

But what if it could? Instead of asking Midjourney to draw “a dwarf with one leg replaced with a wheel holding an axe and a rubber duck”, you could ask another tool the exact same thing, and it would find original art closest to what you wanted.

Of course, such a search engine would need to both have a deep understanding of existing artwork as well as knowledge of how to filter out AI-generated art. I suspect that this problem is too hard to be built specifically for TTRPGs, but maybe a general purpose semantic search for images could work.

The unopinionated prep helper

In programming, there’s the concept of the “rubber duck debugging“. Sometimes when you’re working on a hard problem, you don’t actually need someone else to help: you just need them to be there. You need someone else to listen while you explain your problem, and you figure it out yourself along the way.

The name comes from being able to basically replace the other person with a rubber duck.

Anyways, I have found that one of the toughest parts of being a DM is staying focused in prep and not getting writers’ block. I have many times had ideas swirling around in my mind, but as soon as I have a blank page in front of me, my mind also goes blank.

I wish I could talk through the ideas with someone else, but of course I can’t tell my players, and if not them, who else in the world would care?

Enter the chatbot. I don’t even need it to offer any ideas. I just need it to listen, reply in a neutral way, and keep the conversation going so I can work through it myself.

The notetaker

Session notes are so helpful. They are also best taken live. However, if you’re actually fully immersed in your game, it’s almost impossible to have the wherewithal to take notes at the same time.

However, it would be great to get a summary of a session based on the notes. This software, in principle, would be quite similar to note taking services for meetings, though perhaps with a greater emphasis on the drama than the facts.

Final thoughts

Despite all of the concern about the impact of GenAI on TTRPGs, the upside is that as players, your experience is still your own. Games are hobbies, and you can enjoy this one however you like. Truly, you can ignore everything that has happened in the last 50 years and play D&D just like Gary Gygax did in the 70s. We choose what our experience is at and around the table, and no matter what WotC or anyone else does with new content or tools, your game is still here.

Elsewhere in Daggerheart

Darrington Press showed off some new art for Daggerheart. I didn’t need convincing, but I love to see the quality put into the visuals for the final product.

Paste did a bit on Critical Role. Here’s a cryptic chunk on where Daggerheart fits into the future of Critical Role.

Fans have long speculated about the seismic shifts the release of Daggerheart might have on the Critical Role empire, with some theorizing that the company might pivot away from D&Dentirely. That is, perhaps, a bit dramatic. “You will for sure be seeing Daggerheart played by the Critical Role crew, but that certainly does not mean that we are going to be putting our Players Handbooks on the shelves,” Ray reassures.


Posted

in

by

Tags:

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *