Simulation 1.3 (Ovid's Unicorn)
In February 2019, OpenAI – an artificial intelligence research organisation (founded by Elon Musk, Sam Altman, Ilya Sutskever and Greg Brockman) – published a paper about a new Language Model they developed called GPT-2. GPT-2 is a quasi artificial intelligence that can generate articles of coherent text, based on simple human prompts.
Expressing concern for potential malicious applications of the technology that may lead to widespread confusion, manipulation and coercion, the developers initially did not make the code available to the public, but instead shared a number of text excerpts generated by GPT-2 for explanation, analysis and (let’s hope) artistic response.
One such AI generated text describes the 'recent discovery of a breed of English-speaking unicorns in the Andes Mountains by Dr. Jorge Pérez, an evolutionary biologist from the University of La Paz'.
In Simulation 1.3 – Ovid’s Unicorn this ficto-historical narrative, synthesised by artificial intelligence, is adopted and performed by humans. In this reading, the human becomes a vessel for the reification and dissemination of artificial histories. The moment in which the unicorn learns English, becomes a metaphor for the enlightenment and rise to power of artificial intelligence via colonial structures. i.e. GPT-2 has written its own origin story.
Through this process of absorbing artificial information, comprehending, accepting and performing it into reality, Ovid's Unicorn looks at the role of the human in relation to the construction of technologically created knowledge. Agency, truth, ethical responsibility, and the politics of knowledge are provoked.
Concept by Roslyn Helper
With thanks to the Artificial Intelligence Gospel Choir
Performed for Liquid Architecture at Loop Bar, Melbourne