Abstract

As a survey of algorithmic subject classification and the aesthetics of generative AI, Synthotypes of British and Foreign Ferns is a collection of post-photographic, synthesised specimens, fabricated with generative AI imaging systems.

Seeded in latent space, the 'synthotypes’ are algorithmic depictions of the botanical specimens that originally featured in Anna Atkins and Anne Dixon's study, Cyanotypes of British and Foreign Ferns (1853).

Working from a broad sample of the botanical specimens Atkins and Dixon originally rendered, their Latin names were repurposed and used as the basis of simple text-based 'prompts', which were processed through a text-to-image AI generator. The resulting fabrications were algorithmically upscaled, labelled and catalogued to create a new, post-photographic taxonomy.

Revealing the interpretive nature and procedural flaws inherent in contemporary machine learning and generative AI models, the collective synthotypes reflect the technological shifts, as well as some of the growing concerns with regards to the automation and outsourcing of representation in the age of computational image production.