Artificial Intelligence is having a profound impact on image creation. Its rapid evolution is incredible.
It’s a highly controversial topic for illustrators whose work has been stolen and used to train machines to replicate their art without any compensation.
AI isn’t going away and a crucial aspect of my work is learning and adapting. Every project I’ve worked on recently has involved AI to some extent, sometimes to expedite retouching, sometimes to create entirely new images.
The biggest challenge with AI is control. AI has its own perception of the world. If you want an image of a beautiful woman amidst a fantasy setting, AI can generate thousands of captivating images. However, if you want a specific person in a specific location, the process becomes more intricate especially if it needs to look photo realistic.
AI image generation relies on words, but I’m a visual artist, not a writer. So, after typing in a few words to generate a starting image, I continue to create images until I get close to what I’m after. I then use that image as a reference to create more precise images, modifying it by adding or removing elements, manipulating the shapes of limbs, and so on, an ongoing process of iteration.
The images in this post are from a project with Matt Shelley of Pollitt and Partners, creating 22 enormous window displays for a renowned department store undergoing a transformation into new uses. The new images reflect the statues of the original 1911 building, the largest store in the world when it was built.
Repeated iterations with MidJourney consistently produced images that were often close to, but never quite what was wanted, so I was asked to refine and enhance them using AI, 3D modelling, and illustration, drawing inspiration from Renaissance statues. Coincidentally, I was in Florence during the time of working on some of the illustrations, surrounded by the most wonderful reference material. An article about the project can be seen here
