Statement on Generative AI

Short statement: I am against the wholesale training of copyrighted materials deployed to replace working illustrators, photographers, writers, and other creatives. I believe generating an image or text ‘in the style of’ robs highly trained professionals of their livelihood, reduces the incentive for young creatives to produce something meaningful, and is a corrupting force for society.

I personally am interested in non-determinant art created by early models. I work with the wild and unpredictable models, the early day encapsulations. I don’t want to direct the output to replace, but seek to understand the universe as understood by the machine learning model.

Longer thoughts on the industry trend towards GenAI (published approx. Sept 2022):

I am an long time artist, designer, and now, genAI enthusiast. My career has jumped in several interesting tangents as I adapted to the Desktop Publishing revolution, Photoshop, the Web, mobile phones, and now genAI.

Most artists adapt to new tools. Not all, I have quite a few friends who never grasped the print-to-web shift and they got left behind. I have a friend who is a children’s book illustrator that recently waxed poetic when he went from oil pastels to ipad. He now has access to layers, infinite redos, no smudges, and production-ready art that doesn’t need a drum scanner.

I am constantly delighted by the work I produce using Stable Diffusion. I love the dreamlike quality and appreciate the constant surprise. I don’t want to control the output too much, I want the model to generate an approximation based on the imperfect way our world has been encoded and compressed.

I don’t consider this art, and I don’t consider this non-art, but more of a ’confection’.

There will be a wave of garbage, to be sure, as there was during the DTP revolution. But many will continue to make art of value, in all media, including our latest newcomer. As the Midjourney-plastic slop gets greater, there will be a greater value placed on bespoke, humancrafted art.

Sure, I miss the days of guache (no), rubylith (yes) and stat cameras (hell, yes, I was a good cameraman). I still make art, deciding between ink and paper, digital camera, or film. I also love a session with Stable Diffusion and Automatic1111.

I also agree that some fine stylistic illustrators are being ripped off. I wish there were a good answer, but sometimes humans leverage tools to sleaze.
It’s a new medium, it’s not the end of art. There’s room for all.