On AI.
As an artist and designer it’s easy to scoff at AI image generation, and easier still to fear it. Artificial Intelligence has stolen the show. It’s here and it’s doing things we didn’t think it was going to do. In the art community most seem to share a simmering abhorrence, an anxiety, or anger. People can now leverage AI trained on artists’ work, instead of paying artisits. While I understand and struggle with these feelings myself, I don’t think this is a productive mindset. I don’t think AI is going away, and I believe AI can benefit artists too.
This post will act as an archive of my relationship with AI image generation. I will make noted updates when/if my feelings change. It is also an extended disclaimer of sorts- speaking of which:
Disclaimer:
I use AI generated imagery in my personal projects as a way to create backgrounds and characters for fictional games. This creates context for my UIUX work. I do not, nor will I ever use AI to generate assets used as part of the interface I’m designing. I will never use AI when working on a shipping product. I only use Midjourney AI to simulate the game environment and characters for my personal projects, as a way to create context for my UI work to overlay.
That being said, I would much prefer working with another artist in a way that is mutually beneficial. If you’re an artist who would like to collaborate on future work, please contact me!
Leveraging AI as an Artist.
Controversy aside, it is difficult to argue against the utility of generative ai. It isn’t perfect, but it doesn’t need to be. As an artist I’m not necessarily interested, nor do I really require it to spit out images that are production ready.
Mood boards are a big part of my process. They act as inspiration or guidance for visual tone and aesthetic. I’ll collect all kinds of visual reference, and recently, I’ll even generate some. I’m never taking elements directly from these images, but AI is great at mashing up styles and iterating on an idea until you find something just right. This is also extremely beneficial for any client or stakeholder who may not be able to envision what you describe as ‘Ghibli-meets-Burton-Victorian-Folklore’- for example.
Filling layouts with stock photography can be expensive and the free options are often quite limited. For personal projects or mockups, generative AI can get you what you need relatively quickly, especially if these background elements will be blurred later. AI is quite good at creating generic landscapes or cityscapes that suit the needs of your app or game concept.
When I’m working on a UIUX concept for a personal project I will do some expedited world building via generative AI. For example for my AAA Ninja Turtle concept I used descriptors like ‘Neo New York’ and ‘Cyberpunk sewer’ to start getting excited about the setting. This sort of exploration can get me excited and inspired to continue fleshing out details about the world, story, and finally game mechanics which need a UI to support them. Finally, once I’ve wireframed and prototyped my UIUX, I use AI to generate fake screenshots. I’ll prompt using phrases like, ‘in the style of 3D AAA PS5 game’ as a way to generate backgrounds for the UI to sit on top of. All this to better simulate a finished game.
These have been some of the ways I use AI in my personal projects. I think it’s a fun and quick way to augment the ideation and finalization phases of a concept, when you’re working solo. Working with a creative team brings a lot of advantages that AI cannot emulate. There is no real back-and-forth in terms of idea development, it isn’t critical, it isn’t considering things outside the prompt or bringing experience or perspective to creating things or solving problems. AI empowers the solo creative, it can augment a process in some ways, but it cannot replace a creative team.
