Indie game development has always been about doing more with less. AI image generation is turning "less" into "remarkably enough."
The Indie Problem
Making a game requires:
- Game design
- Programming
- Audio/music
- Writing/story
- Art
A solo developer might be great at one or two of these. The others become blockers. This is why most indie games are either 2D pixel art (easier to make) or take years to complete.
AI changes this calculus.
What's Actually Happening
Concept art. The fastest application. Need to visualize a character before committing to the design? Generate 20 options in an afternoon.
Prototyping. Instead of placeholder art, devs can generate something that feels closer to the final look. This helps communicate ideas to collaborators or publishers.
Asset variation. Generating multiple versions of props, backgrounds, environmental elements. Creates variety without hand-crafting every variation.
Marketing materials. Generating screenshots, store page images, social media content. This takes time that could go into making the game.
UI elements. Icons, buttons, menu backgrounds — not glamorous, but necessary.
Real Examples
Visual novels. These are particularly suited to AI since they heavily feature character portraits and backgrounds. AI lets solo devs compete with teams.
Roguelikes/cycroguelikes. Procedural generation + AI art = massive content variety with less manual creation.
Adventure games. AI can help visualize scenes, characters, and objects. The AI doesn't replace the artist's vision — it helps execute it faster.
Game jam projects. At game jams (intensive short game creation events), AI is a game-changer. A game that would require an artist now can be made by a programmer alone.
What It Can't Do (Yet)
Consistency. The biggest problem. Characters generated separately might not look like the same person. This is crucial for games where you're looking at the same character for 20 hours.
Animation. Still requires frame-by-frame work or sprite sheets. AI can generate images, not animations.
Specific compositions. If you need exactly this character in exactly this pose in exactly this context, AI might take many tries — or never quite nail it.
Integration. The AI output isn't game-ready. It needs to be imported, optimized, integrated. That work still requires human skill.
The Controversy
Not everyone is excited:
Professional artists worry about displacement. If indie devs don't hire artists, where does that leave professional game artists?
Quality concerns. AI art has a certain "look." Games might start feeling generic if everyone uses the same tools.
Ethics questions. Whose style is being learned? Are we comfortable with that?
These are real concerns. The answers aren't simple.
A Possible Middle Path
Maybe the healthy relationship is:
- Use AI for prototyping, iteration, exploration
- Hire human artists for final, polished assets
- Be transparent about what's AI-generated
- Credit AI tools alongside human work
- Support artists whose work helped train the models
This isn't free of problems, but it's a start.
The Bigger Picture
What excites me: more games will exist. Games that would never have been made because the creator couldn't afford art now can be made.
Not every game needs to be a AAA masterpiece. Some games just need to exist. AI lets more of those games happen.
That's worth something.
The barrier to entry keeps lowering. What game will you make at ArtFelt?