Working with AI art sometimes feels like gazing into a pond of water. In a still pond, we see a clear, gently blurred version of ourselves staring back at us, while turbulent waters return rippled echoes of our shape. These changing reflections of ourselves are similar to images generated from data, which can be hyperrealistic depictions of the original, or images that are surreal and barely recognizable, as errors creep in. This exhibition looks at the way artists represent nature and humans, using the latest machine learning technologies to conjure seascapes, imaginary marine creatures, autumn leaves, and musings on the human form. Masks and faces akin to the grotesque rear their heads, reminding us of the early errors of these technologies and the eternal human fascination with depicting ourselves. These artists present AI as a tool to craft new reflections on perennial subjects, as well as a creative partner, conjuring new landscapes, creatures, and forms. Luba Elliott

Feral File  Reflections in the Water

&ldquo;Leaves of Manifold, of Red and Gold&rdquo; by Helena Sarin is a piece that focuses on autumn tree leaves, the rich combinations of their color hues, and their seasonal omnipresence. Leaves appear and disappear one by one, presenting a fashion parade to the viewer, the detailed veins and complex color transitions providing a feast for the visual senses. At times the shapes seem to belong to unfamiliar trees, or the leaf stalk may be separated. As is typical in Helena Sarin&rsquo;s practice, these images have been generated by a GAN trained on the artist&rsquo;s own photographs and sketches. Then, Sarin reviews the generated images and manually curates the final selection, acting by the principle &ldquo;It&rsquo;s not the frames you play, it&rsquo;s the frames you don&rsquo;t play,&rdquo; to paraphrase Miles Davis.