Digital costume designs for Influenza 1918

Inspired by the successful way I used the LLM-driven aspect of generative AI to plan a workshop schedule, I decided to import some of my most successful Copilot-generated costume images into Canva and see how they would look formatted as costume design renderings. 

I realized that I had been letting the AI hype gaslight me into thinking that I could expect the technology to completely “do my job,” instead of create raw material or a draft that I could then manipulate using other tools.

Be advised that I am not super proficient with the Canva image-editing app, so I’m sure a person more proficient in digital rendering tools could do more with this idea, but I was able to eliminate busy backgrounds and that one weird uniform button that was pissing me off. 

I envisioned these designs were for an opera about the influenza pandemic of 1918 and its impact on Africa. Which, I realized I knew nothing about so I read about it on this website. An interesting and tragic research side-quest, as it were.

Of course, if this were an actual opera, an actual production, I would hope that the costume designer would be a talented designer of color and not me. So I’m treating this kind of like a class assignment with the instruction that I’m trying to create some initial renderings I could take into a first creative team meeting with the director & the other designers.

All this is entirely backwards from how something this like this would happen in actuality, and presumably the composer and librettist of the opera would determine at least the location (because Africa is a big continent& not a monolith) and I would be able to do research and then draw these costume designs or have that information to impact the prompt language for how the renderings were generated.

But, i digress. II was able to create a rendering template in Canva and import the Copilot artwork, clean things up and produce these just like I would do if I had sketched these. When I was working as a costume designer, it was at a point in the progress of technology such that I often drew my renderings by hand on paper, scanned them, and imported them into a template just like this. In creating that template I put thought into font choice and layout of the rendering for concise communication to the costume crew, my fellow designers on the creative team, and the director.

I also added a little AI graphic in the corner of each to communicate clearly that the generation of these designs was done with the help of a diffusion model.

Here are my results: