Rainbow Wolfes, AI Artist

Our AI virtual Artist in Residence is more virtual than most this month, through the use of artificial intelligence to create a series of images. Through a process known as stable diffusion, a text description – ‘Rainbow Wolves’ – was converted into a numerical representation called an embedding, which could then be understood by the neural network. This is then generated as an image, through a ‘diffusion’ process of blurring and unblurring. The final result is one of uncanniness, an image with no physical link to the real world but which has been composed completely from learnt behaviors.

The improvement of AI is now moving at a pace that is difficult to keep track of, with serious implications in the world of art. It can be argued that the creative process involved in developing the AI model and training it on a dataset of existing artwork is similar to the creative process involved in traditional art-making, paving the way towards a future in which AI can create images indistinguishable from human artworks.