< Back to Blog

Game Design with AI

Welcome to the inaugural post of our new series where we discuss artificial intelligence and its impact on game development at large and at our studio! In this game design edition, we include the insights and experiences of Filament Games’ Chief Creative Officer, Dan Norton, as he navigates the intersection of AI and game design.

via Giphy

Want more posts like this one on your social media feed? Follow us on X and Facebook for daily articles on game-based learning, technology, education, and more. 

Not to get TOO meta about this, but I’m writing this essay, in first person, entirely with just me and my meaty fingers. No feather quill with hand-ground ink, no parchment, but still, I’m the one pressing the keys on the keyboard to make the words in the order I want.

That’s because I don’t think I can live with the irony of having my personal philosophy about AI being generated (or in this case, even shaped) by an AI. If you’ve worked with enough AI text, you can catch the “AI vibe” from reading it; an obsession with balanced statements, neutral positioning, personality only inserted when it’s ultimately harmless…it’s got a feel. And I don’t want that feel anywhere near my own statements on the technology itself. Let’s keep it spicy!

Right now, AI is sweeping through every creative field. And in those fields, there’s apprehension, excitement, anger and confusion. Everyone who works in a field that involves thinking of things that take the shape of words, images or videos is now wondering whether they’ll have a role in the future, and indeed if what they do ever mattered in the first place.

Of course it matters. The only reason AI is able to make things that roughly feel like an aesthetic experience made by a person is that they’re pouring your cup of art from a gigantic smoothie, blended from millions upon millions of prior creative works, made by countless people over years and years of effort. Almost none of them asked to be part of the experiment, and almost none of them will receive any attribution or reward for their tiny drop, but there they are. AI is nothing without them, and soon AI’s contributions into the creative field mean we’ll just be pouring the smoothie back into the blender in larger and larger proportions.

But that’s a lot of fairly grumpy words for someone who’s using AI on a regular basis. I use it, the design team uses it, the art teams use it, the marketing team uses it, the programmers use it. Each team finds some genuinely cool new application every week or two, and it’s…exciting! The design team uses it to help with arcane spreadsheet manipulation, or placeholder content generation, or even just as a sounding board for generating cheap ideas.

And that’s all great! But how do I reconcile these two stances? It’s not perfect, but I’ve got a general rule I’m running, and it’s holding up…so far.

I think for me, as of right now, the simple line I’ve drawn, is that AI is great for generating ANY type of content, for ANY sort of scenario, as long as…that content doesn’t matter. And by “matters” I mean isn’t an important part of the creative communication between artist and consumer (or, developer and player, as the case may be over here).

I think for many people, art, music, food, comics, games are an experience where we take in a medium as an expression of thought from another person or group of people. We read a book, and consider what the author meant behind a choice they made. We watch a horror movie, and delight in the practical effects that we know someone hand crafted. We listen to a song and wonder what types of drugs, in which combination, could induce someone to make it. The connection between creator and consumer is what makes art…matter.

It’s not an uncommon critique, and certainly must have been even a prevalent one, when photography first burst into the art scene, that it was an automation of the artist’s expression. Gone was the handcraft of paint mixing, the keen eye for light and form…and in its place is some person who just points a box at something, and it generates a rendering that looks like it came right out of someone’s eyeballs into yours.

But, obviously, people found there was still quite a bit of “art” still happening. There’s the interesting technical decisions, of course, but frame composition, light, color, form, narrative…those things are all still there, and are choices that the photographer can make, and are things that the consumer of that photo can still consider as a connected experience between them and the artist.

And I don’t think the crossroads we’re at now is remarkably different- maybe there’s more crossroads at once, and the decisions are certainly coming at a pace that’s hard to keep up with, but it’s essentially still the same questions:

What parts of this medium are the things that the artist is using to express an idea? What is that idea? And what do I think of both that idea, and the expression of it?

If someone types “dragonborn lady riding a skateboard”, and sorts through 75 reconstituted renderings of a scene, tweaking just how dragonish that lady will be, or how sick of a skateboard they have, I personally don’t think there’s really any art happening. Sure they’re making choices based on each pass, but that’s what I do on Amazon, and I don’t expect any accolades for picking out the best toilet brush as an artistic expression.

Essentially, every decision that matters is one that the “prompt engineer” has outsourced to the engine. All the things that I find interesting about an image are ones that this person has skipped. The only idea that they’ve created entirely from themselves is “dragonborn lady riding a skateboard”, and every other choice has been based on micro-robbing hundreds of thousands other images of skateboards and dragons.

But let’s say instead you have a background art environment for a dialogue interface. There is a forest in the background. Honestly, if that forest is AI generated, it doesn’t detract from that experience as a user for me because…the background doesn’t really matter much. It’s fine. I’d rather the artists spend their time on pieces that more effectively communicate their vision, if they can get some totally acceptable trees just whipped up real quick.

At any rate, that’s my personal philosophy. And as Chief Creative Officer, I have the luxury of dictating my philosophy as if it’s policy. So, if this rambling is be transmuted into some sort of mission statement, then it’s this: At Filament, art (code, sound, image, text, video!) will be made by artists for every part of expression that matters– matters to the team, matters to our clients, and matters to our players. We’re thrilled about exploring the benefits of AI to empower our teams to make art better, faster, and more iteratively, but making a game is a creative endeavor, and without a creator, a creative endeavor is impossible.

By approaching AI with a focus on enhancing, rather than replacing, human creativity, game designers can harness its potential to create more high-quality gaming experiences. Stay tuned for more posts in this series, where we’ll delve deeper into the practical applications, challenges, and future possibilities of AI in various areas of game development. 

Ready to create a game-based learning project for your organization, school, or workplace? Contact us today to discover how our innovative approach to game design can combine with your ideas to create something truly impactful. 

More game design insights:

© 2024 Filament games. All rights reserved.