Demolish Monsters on the Rocks: Prompting Through an Ensemble

bruce gilchrist

As interaction with corporate artificial intelligence increasingly becomes a precondition for contemporary life, artists need to see beyond Generative AI (GenAI) technology as a discrete tool that makes generic products. Instead, they can imagine combinatorial approaches and conceptual frameworks for AI-enabled artworks. Through my practice-based research, the act of prompting multimodal GenAI models has been informed by comprehending an assemblage as a “framework of instruction” held together through poetic alliances, within which the output from one component feeds the process of another. Practical experiments explored an interrelation of body, text, and predictive technology, where an algorithmic prediction of human action conjured “biometric poetry” that was used to stimulate a language model. Working with archival film footage and digital puppets animated with motion-capture files gave rise to the idea of a camera’s field of view – with its bounded contents acting like a key – eliciting value from a language model in a novel form of story making. Potential erroneous inferences were perceived as a new form of chance operation and a characteristic of algorithmic remix as defined by Steve F. Anderson. This method has been further developed in a project that combines performance, waste material, object recognition, and a language model to explore how the manipulation of garbage can be rationalised by a machine to produce poetic texts as a commentary to action portrayed on a screen.

09 Distilling Literary Corpora

Three of the practical projects that will be outlined in the following sections, Synset_Gloss, Idling-in-the-Unreal, and No_m_oN, employ a language model trained exclusively on the works of Michel Foucault as part of the ensemble.

Contents Back | Next

Back | Next

Foucault’s seminal works Discipline and Punish (1975) and The Birth of the Clinic (1963) were two of seventeen volumes used to train the language model employed in practical projects. In his essay “The Theory of Vibe”, Peli Grietzer (2017) explores how AI can process literary corpora within a mathematical philosophy of literature. He proposes that training an AI language model on a single author’s body of work enables the distillation of that author’s “vibe” into a mathematical abstraction, independent of any specific instance of their style. Following Grietzer, a language model based on Foucault forms a significant part of the aforementioned projects. Foucault’s philosophy of power is crucial in understanding the evolution of what could be termed info-power – the control and influence exerted through data analytics, social media, and continuous algorithmic assessment. Since his death in 1984, this new form of power has become one of the most significant forces shaping contemporary society. Foucault explored how surveillance and governance mechanisms such as panopticism and biopolitics are central to the exercise of power in modernity, particularly through the categorisation, normalisation, and control of bodies. Biometric technologies, which quantify physical characteristics like facial features and fingerprints for identification and surveillance purposes, reflect the extension of disciplinary mechanisms into the digital realm. Foucault’s analysis of how bodies are subjected to power and knowledge systems is essential for critiquing contemporary surveillance technologies, which now operate at the intersection of embodied data and state or corporate control.

Figure 12. Oeuvre, 2021. A corpus of Michel Foucault visualised using Word2Vec Word Embeddings and t-SNE.

The 17 volumes constituting the corpus used to train the language model for Synset_Gloss, Idling-in-the-Unreal, and No_m_oN have been visualised using Word2Vec Word Embeddings and t-SNE (fig. 12). T-SNE is an ML algorithm for data visualisation that performs a dimensionality reduction technique, mapping multidimensional data to two or more dimensions – a new data representation that makes it human-readable, while preserving neighbourhood relations.³

Tip

Tip