Corpus (2020)
“what are the humans doing and what might they do next”
In both projects Synset_Gloss and Idling-in-the-Unreal an algorithmic “prediction” of human action conjured biometric poetry that was used to stimulate a language model to remix 17 volumes of Michel Foucault’s oeuvre. The initial algorithmic analysis asked “what are the humans doing and what might they do next”1 and is analogous to a new form of surveillance power. Foucault is a philosopher of power, and the power of data (the info-power of social media, data analytics and continuous algorithmic assessment) is arguably the most significant kind of power that has emerged since his death in 1984. This becomes more acute as artificial intelligence becomes increasingly coupled with surveillance technology, contributing to an epoch that already feels crushed and exhausted; an age of generative AI and synthetic knowledge. The volumes constituting the corpus used to train the language model for both Synset_Gloss and Idling-in-the-Unreal have been visualised below using Word2Vec Word Embeddings and t-SNE︎︎︎
1 In 2020 the mission statement of the Perception Team for Google AIs AVA (Atomic Visual Actions) was to imbue computers with “social visual intelligence”: what are humans doing, what might they do next, what are they trying to achieve? https://research.google.com/ava/
