Ben Millar Cole is a traditionally trained artist who began experimenting with artificial intelligence in 2017. Interested in the creative limitations of synthetically guided creativity, Millar Cole created his own Recurrent Neural Network (RNN), which could generate audio/visual outputs based on datasets. These datasets are sometimes publicly available, for example Millar Cole used an online BBC sound archive of over 16,000 recordings, all prosaically named after what they’re describing (Shouts of encouragement at a Wrestling Match or Large Bird Taking Off. This dataset provided him with the raw material to play with how image and language can signify each other under the influence of a neural-like AI system.
Millar Cole feeds his RNN the BBC’s descriptive text and it engages in a game of probability to create its own outputs with surprising results. ‘Hoods Pausing, Tender Roles’, ‘River Pig loading’, and ‘Comedy Gold, Before Scream’, are a handful of its creations. These nonsensical failures produced by early RNNs are the happenstance, poetic absurdities Millar Cole seeks to build on. He then playfully creates images that visually articulate an interpretation of each output’s linguistic meaning. In other words, he moves from sound, to text, to AI, to text, to image.
As the technology improves, he is engaging in an ever-challenging battle of creative authorship with the machine. Recently, a new breed of AI image-generators have broadened the possibilities of magical thinking and image making. These expand our understanding of how creativity can be synthesised, amplified and manipulated – and further complicate the notion of photographic veracity.
"The work of Ben Millar Cole combines unexpected phrases generated by older recurrent neural network-based (RNN) systems from the mid-2010s, which generate text one letter at a time, with the latest image generation possibilities of Midjourney, bridging two different eras of AI art: one that glorifies the creativity of the limitations and errors of the machine and the other that is much closer to perfection, but still with the existence of the uncanny valley. This is underscored by the close relationship between titles – put together they could form a poem – and the image, guiding the viewer’s interpretation of the human-like figures, themselves almost a remnant of the early GAN era, in which images with mislocated limbs were much more common."
Excerpt from Post-Photography: The Uncanny Valley, an essay by Luba Elliott