ADG:   Machine Auguries: Toledo is a 12-minute orchestration of a dawn chorus, gradually taken over by artificial birds, created using AI. I used a generative adversarial network (GAN), which is a pair of neural networks – a generator and a discriminator – that, like birds in conversation, are performing a call and response. Having been trained on datasets of things that already exist, the GAN can then produce new images or sounds. It’s the technology behind deep fakes and DALL-E. In this case, my GAN trained on 100,000 field recordings of local and migratory bird species iconic to the Toledo dawn, provided by the Macaulay Library at the Cornell Lab of Ornithology.

Visitors enter the gallery in the gloaming blue light of the pre-dawn and sit in a clearing under a vast artificial sky. An American robin punctures the hum of crickets as it utters the first call of the dawn. It sings again. Then a machine sings back from across the clearing. This first artificial bird is from very early in the GAN process, so it is glitchy. Over time, the calls increase in fidelity as more birds join, from the northern cardinal to the black-capped chickadee. As the chorus swells, the sky washes to a silvery blue and then erupts into the peach of the breaking dawn. Deep machine sounds swirl overhead across 24 speakers. Slowly, the calls thin as the sky rolls into the cool light of day. By the end, it’s unclear whether what we are listening to is real, but it is all artificial. The last few birds sing, and then it is silent. We are indoors, in the absence of nature, taking time to listen to an unnatural reconstruction of the unreplicable natural world outside.

The dawn chorus peaks in the spring and early summer as birds mark territory and find mates. We think of habitat loss and pollution as the reason for dangerously declining bird populations, but human-made noise and light pollution have insidious effects as well. Artificial light confuses birds: how do you sing at dawn if the day lasts all night? Noise pollution is also making birds sing earlier, louder and at a higher pitch – if they can. Birds play a vital role in functioning ecosystems: I want to focus on the beauty – and the fragility – of their song.

Machine Auguries: Toledo is the second in a series; the first iteration was commissioned by Somerset House recreating a UK chorus in 2019. I’m increasingly making series of site-specific works, like this and Pollinator Pathmaker, to bring attention to the uniqueness of local ecosystems while showing their interconnectedness in a larger system.

TW:   What places or books do you find the most inspiration from?