Is Every DM Really Using AI?

No. Not every DM is using AI. But when I went to DMU (Dungeon Master University) in Atlanta this January, a whole lot of them were.

Including multiple instructors. Some of whom are world-famous DMs.

What I Saw at DMU

DMU is exactly what it sounds like: a conference where Dungeon Masters learn to be better Dungeon Masters. Sessions on encounter design, narrative pacing, player engagement, world-building. The kind of event that makes non-D&D people question your hobbies and D&D people ask for the registration link.

I went expecting to learn new techniques for running my Eberron campaign. I did. But the thing I didn’t expect was how many DMs were already using generative AI as part of their workflow. Not hypothetically. Not “I’ve been meaning to try it.” Actively, regularly, in their actual games.

Some mentioned it openly in their sessions. Others told me privately, usually followed by something like “please don’t post this on BlueSky.” The tabletop community has strong opinions about AI, and not all of those opinions are friendly. Getting publicly identified as a DM who uses AI tools can get you dogpiled faster than a level 1 character walking into a dragon’s lair.

What DMs Are Actually Using AI For

The use cases were practical, not philosophical. Nobody was asking ChatGPT to run their entire campaign. They were using AI to solve specific problems that eat a DM’s prep time:

• Character and setting art. Instead of “borrowing” images from Google (which is what most DMs have done for decades, and let’s be honest about that), they’re generating custom character portraits and location art. The wizard looks like their wizard. The tavern looks like their tavern. No more pulling fan art from DeviantArt and hoping the original artist never finds your campaign Pinterest board.

• Battle maps. Same principle. Instead of downloading someone else’s dungeon map and pretending it fits your layout, you generate one that matches the encounter you designed. Custom dimensions, custom terrain, custom lighting. The map finally matches the story instead of the other way around.

• Dialogue and speech writing. This one surprised me the most. AI is genuinely good at period-appropriate speech patterns. When the noble addresses their people, the proclamation sounds like a proclamation. When the dwarven merchant haggles, the phrasing has texture. Writing 15 distinct NPC voices for a single session is exhausting. AI handles it in minutes, and the DM edits for tone and personality. (My improvised noble speeches sound less like a proclamation and more like a CEO who vaguely remembers his Shakespeare. AI is an improvement.)

• Idea generation. “I need a plot hook involving a cursed lighthouse and a merchant guild.” Try getting that from a random encounter table. AI gives you five options in 30 seconds, three of which are usable and one of which is genuinely better than what you would have written at midnight the night before the session.

The common thread: DMs aren’t replacing themselves with AI. They’re using it to do the parts of prep that are tedious so they can spend more time on the parts that are creative. That distinction matters more than most people in the debate seem to realize.

Then Ryan Dancey Got Fired

On February 19, Ryan Dancey was terminated as COO of Alderac Entertainment Group. His offense: saying AI could “come up with ideas as good as” some of AEG’s published game designs.

If you don’t know Ryan Dancey, you should. He’s the person who created the Open Gaming License, the legal framework that arguably saved D&D in the early 2000s and spawned an entire ecosystem of third-party content. He’s one of the most important figures in tabletop gaming history.

What Dancey said was clumsy. What he meant was closer to what every DM at DMU was already doing: AI is a useful tool for generating raw material that a human then shapes into something worth playing. He wasn’t saying AI should replace game designers. He was observing that the idea-generation step of design is no longer exclusively human territory.

The tabletop community heard “AI is as good as our designers” and reacted accordingly. AEG fired him within days.

The Nuance Nobody Wants to Have

Here’s the thing Dancey couldn’t articulate in time: there’s a meaningful difference between AI generating ideas and AI replacing the people who turn ideas into finished games.

A DM using AI to draft NPC dialogue isn’t replacing a writer. They’re a hobbyist who was never going to hire a writer in the first place. They were going to spend 45 minutes writing mediocre dialogue themselves, or they were going to skip it entirely and improvise badly at the table. The DM who generates a battle map isn’t taking work from a cartographer. They were going to use a free map from Reddit or draw something on a whiteboard. These are people enhancing a hobby, not eliminating jobs.

A game publisher using AI to replace their design team is a different conversation entirely. Dancey wasn’t advocating for that, but his words left enough room for people to assume the worst. In a community already anxious about AI, that gap was all it took.

The DMs at DMU understand the difference intuitively. They use AI because it makes their games better for their players. They’re not taking a job from anyone. They’re spending less time on prep they don’t enjoy so they can spend more time on the creative work they love.

But the fear is real, and it’s not irrational. The same technology that helps a DM write better NPC dialogue could help a publisher cut their writing staff. The difference is intent and context, which are exactly the things that get lost in a BlueSky thread at 11 PM.

What I Think

I’m building an AI-driven DM’s screen. I’ve been a DM for years. I use AI in my campaign prep. So I’m not a neutral observer here (and I’d rather be honest about that than pretend otherwise).

I also think the tabletop community’s instinct to protect creators is correct, even when it gets misdirected at individual people like Dancey. The answer isn’t “AI bad” or “AI good.” The answer is: AI is a tool. DMs have been using every tool available to make their games better since Gary Gygax was drawing dungeons on graph paper. This one just happens to be controversial.

The DMs at DMU aren’t waiting for the community to reach consensus. They’re too busy prepping their next session.

(I’ll allow it.)