Photos by Michael Hull, courtesy of Times Square Arts
“Critically Extant” is a project that explores just how little we know about the natural world by testing the limits of the data openly available to us in our digital lives.
To achieve this, AI algorithms were trained on millions of open source images of nature and some ten thousand species. The resulting models were then used to generate visual representations of species that are critically endangered, yet have little or no online presence specially on social media The goal of this was to not only trace the edges of our knowledge, but to also explore how we can create feedback loops in the digital that can be positive for the natural world.
The project was inaugurated as an Instagram exhibition, exploring how the pieces can become part of our daily digital intake of content as a means of creating awareness and potentially engagement on behalf of the species shared. Naturally, as the data available to us represents but a partial fraction of the real number of species currently estimated as known to us, the pieces in this series show animated specimens that bear some, little, or even no resemblance to the species they are meant to depict.
This underlines the difficulties we face in shifting also our digital spaces towards more balanced representation, but it should be grounds for agency too: as we can all create and contribute both physically and digitally and as such can actively work to form new feedback loops that can help bring the critically endangered species into our daily lives in order to get to find ways to care for them?
This project was made as part of the 'Meta AI Artists in Residence' program.
2023 The Photographers' Gallery (UK) (link) 2023 POSTNATURE - Casa Hoffmann (CO) (link) 2022 Arebyte 'Futures Past' (UK) (link) 2022 Times Square Midnight Moments (US) (link) 2022 "Latin American Futurity" at EP7 Paris (link)
Diagram of process: once having a "nearest neighbour" within the tree of known data, it's corresponding class within the latent space of the neural network was used as a starting point to try an "reconstruct" the critically endangered specimen in question.