Room Escape Artist

“What are your thoughts on artificial intelligence and escape rooms?” This is by far the most common question that I get asked in 2026.

In spite of that fact that I am a tech consultant and Lisa has been working in machine learning and generative AI long before most people knew what ChatGPT was, we’re not especially big on hot takes. So here’s my ice cold take on a topic that most everyone already has some opinion about.

What am I not getting into?

Before we dive deeply into this topic, I want to be very clear that I am not going to dig into the following topics in a meaningful way:

• Environmental Impacts
• Valuations & Finance
• Labor Effects
• Societal Shifts
• Supply Chain Shortages
• Security Concerns
• Governmental Policy

I have opinions on this stuff. You almost certainly do too. And there is more substantial analysis available by people who specialize in all of these things (and your idiot uncle on Facebook).

I’m also not getting into our own use of these technologies as I have already covered them in our AI Policy post. Nothing has changed in our approach since I wrote this in June 2025.

I’m the escape room guy, and I am going to focus in on how these technologies can impact escape rooms and immersive games.

The last thing that I am not here to discuss is whether the technology has uses. It very clearly exists and can do stuff. The questions that I want to address are:

• How useful is AI for escape rooms and immersive games?
• What risks am I assuming when I use it?

What do I mean when I say AI?

“Artificial Intelligence” is marketing fluff language used to describe a collection of technologies. Some of these technologies have been in circulation for many years doing things like catching credit card fraud or facilitating medical research. The technology that most people think of as “AI” is large language models (LLMs) and transformer models… The stuff commonly associated with Claude, ChatGPT, and the ever-growing list of aspirants to the throne of the “everything machine.”

“AI” is a sexier, more sci-fi term that marketers, fundraisers, and lobbyists like to use because it means nothing, but decades of media has conditioned us to feel that it’s important.

So let’s revise the question to what people are really asking: “What are your thoughts on LLMs and escape rooms and immersive games?”

What are your thoughts on LLMs and escape rooms and immersive games?

Escape rooms are at their best when creators recognize that they are crafting human experience. There are a lot of options for entertainment available to your customers, and they have made a conscious decision to do something in real life. It’s on escape room creators to honor that.

Every one of your guests could choose to chat with a chatbot whenever they want. It’s not like this is some rare technology that no one knows how to find. Everyone and their grandmother has access to LLMs. There is absolutely nothing special or novel about interacting with them. Your customers are coming to you for a novel experience.

If you want to put this technology into your game, do it in a novel and human way that doesn’t feel like everything else that they have access to (or are being forced to use for work/ school). The technology must serve the experience. Escape room technology is at its best when it simply feels magical.

I’m not saying that all use of LLMs in escape rooms is bad. And reflexive “AI = BAD” pieces are just as silly as,pieces espousing that “AI is the future and all will be assimilated.”

You can use this technology well (and it can be kind of bonkers for writing code). Most do an uncomfortably bad job.

If you want a simple rule to follow:

If it looks and feels like an LLM did the work, you probably made a big mistake.

What type of things can AI do for escape room creators?

Like with most technology in escape rooms, AI works best when it’s invisible. The best uses of AI in escape rooms that we’ve encountered are the ones we didn’t realize were there.

We haven’t seen very many successful front-of-house uses yet. That is to say, it hasn’t proven to be better at writing puzzles or gameplay, or enhancing scenery. There are specific narrow uses cases in these domains where AI can be helpful. But it isn’t going to (successfully) create in full.

It’s likely more useful in back-of-house cases. It might work for you for some coding. It can supercharge work you’re doing in spreadsheets. It might play a discrete role in various business systems, or in your marketing. It can have utility for you and your business.

What risks am I assuming when I use it?

At the customer level, you’re risking your reputation and customer relationship when you put LLM content into your games, especially without cleaning it up. If imagery, video, or dialog feels like slop, then its slop… and that’s true even if it was made by a human whose talents didn’t live up to the occasion.

The other risk you’re assuming is that an LLM is a tool that costs you a variable amount of to produce work of variable quality.

This technology is not cheap. Its prices have been artificially depressed in a transparent attempt to build hype, marketshare, and monopolies. The risk of getting hooked on this technology and building too many dependencies on it are that:

You may not be able to afford it in the future.

Ask yourself these questions before becoming dependent upon an LLM:

• Is there a better or more reliable solution that doesn’t rely on an LLM?
• Is it ok (or even good) if the output is heavily variable?
• Can I ensure that this is up to my quality standards?
• If customers are interacting with the output, is it facilitating a human moment?
• What will I do if this technology goes away, becomes too expensive, or simply doesn’t work the way that I expected it to in the future with no notice?

From my perspective, I fully accept that this technology exists and it will continue to exist and impact the world.

What I reject is that it will do absolutely everything that the boosters claim. The same people who are telling you that you must assimilate AI into everything are also the exact same people who told you that crypto was going to reinvent economics as we knew it, that JPGs of monkeys were going to be worth inconceivable amounts of money, and that we would all own property in the metaverse.

All of these technologies are real and have uses… but they never play out the way that the hype men tell you they will. Find the right uses, maintain your humanity, and have plans for when some of these companies jack up their prices, change their products, or wake up one morning and realize that they are out of money to shovel into their money burning machine.

And if things change, we’ll update our perspective.

On the subject of money…

Thank you to our Patreon supporters for helping to ensure that the humans over here can continue to do human stuff.

One thing that LLMs do very well is damage trust in the written word and the people who make it. We, as always, remain committed to that craft, and have endless thanks to the folks who make it possible to keep going.

Special shout out to those who joined this month: Anonymous, Galya Titova, Marianne, Lesly Meunier, Mary Robinette Kowal, DUSTIN B WOODARD, Gove, Jason Siner, and Oni Hunters

To those of you reading who have the means, we hope you’ll consider joining this wonderful group of supporters.

Support Room Escape Artist’s Mission

There are lots of ways to support Room Escape Artist, like buying from Amazon or Etsy after clicking into the links included in this post or backing us on Patreon.

The money that we make from these helps us to grow the site and continue to add more value to the community that we love so much.