My dumbass went for a new AMD graphic card earlier this year so I am pretty much locked out of local generation for now. I had a Nividia before. In terms of characters I very much agree, there is no large difficulty in creating something at least passable with AI, but I am more or less thinking about all the other assets. Like menues, etc. which I don't know how to generate with AI if that is even possible realistically.
yes, UI assets are possible with AI.
it's not just for generating characters, also icons and backgrounds.
for panels, boxes, you can just use rectangular boxes with transparency and should be enough approaches to make it look decent
still, photoshop experience for be of help.
see in my interface, it's mostly the same generate icons, or character image
but presented in a way that you dont need extra "designs"
you can just focus on how you present it, not on what decorations you have to make
edit:
You must be registered to see the links
edit: for loras that are sd1.5, generate with 1.5, then img2img with illustrious or the preferred model