Prism (2024)

“AI is just us. AI is human labor obfuscated through a terminology called AI” – Holly Herndon

Typically, when using text based generative AI, prompting is considered an art form in itself, with the specific sequence of words constructing exactly what the artist intends. This work reverses this dynamic, with the prompts open to a wide range of interpretation and response. The AI model provides a non-human perspective on human data and experience – a diffracted view, like a prism that splits the light into reconfigurations.

Prism is a collection of AI generated audio / visual fragments created using text-based generative AI models responding to a series of minimalistic prompts: “memory”, “together”, “body”, “home”, “tomorrow”, “commute”, “nightclub”, “us”, “together”, “body”, “home”, “fear”, “waiting room”, “commute” and “waiting for”.

Through these open ended and abstract prompts the AI reveals not just information about what it’s been trained on – it reveals a deeper materiality to the network. The imperfections and glitches, often referred to as “hallucinations”, in this case are instructive – revealing the mathematical structures and fractal-like details within the high-dimensional space which stores the model’s perception of the world.

AI is an incredibly fast moving and important technology, beyond the obvious significance in capital terms. It is inherently “us”, as Herndon says above – it is our data, our patterns that it diffracts. It is a form of pattern recognition and replication, but then as the patterns get more detailed the more philosophical question arises – what does it mean to be human? And how clear a pattern do we need to see before we might recognize ourselves.


Goldsmiths University, London, December 2023
“Navigating the Not Yet” – Public Lecture @ Tickbird and Rhino, Jan 2024