This thesis investigates the materialism of distributed forms of synthetic intelligence and characterises the paradigm shift implied by the proliferation of AI technologies. It examines the ontological politics of AI’s analytical, generative, and learning processes, exposing the metaphysical burdens embedded in AI architectures and infrastructures. The thesis analyses the transformed economy of meaning production and distribution in contemporary knowledge systems, which differ not only from the pre-digital past but also from earlier forms of algorithmic circulation. Without ascribing inherent culpability to AI for exclusion or social bias, it explores how AI can participate in the reproduction of fascistic logic, focusing on AI-generated simulacra, their circulation dynamics, and their topological transformations. The synthetic structures analysed are approached as complex, multimodal, distributed systems far from equilibrium, studied through complexity science and the tensions between old and new materialist philosophies. The philosophical investigation is grounded in art practice, involving the poietic manipulation of AI systems and experimental engagements with generative algorithms. Chapter 1: The Book of Meat vivisects the materiality of synthetic circulations, exploring how fascist seduction and the clustering of exclusionary content operate within distributed synthetic systems. It analyses forms of cohesion that emerge across multiple simultaneous logics, enabling rapid shifts, cascading effects, and break-offs of meaning. Moving away from post-psychoanalytic and instrumental interpretations of fascist mass psychology, it instead emphasises the sensuous dimensions of violence and the mimetic pleasures that emerge through the circulation of algorithmically generated media. Chapter 2: The Book of Clouds proposes an original reformulation of mood as a material orientation of the political, always situated in the systemic present, rather than expressible only through future or past modalities (neither foretaste nor aftertaste). Building on Heidegger’s “mattering to,” Golding’s “return of difference,” and Sharpe’s Wake, it presents mood as a plural “turning”: an emergent, distributed mode of motion comparable to a tensor field. This redefinition enables the identification of new emergent forms through which exclusionary atmospheres are captured and perpetuated across the architectures and outputs of generative AIs, particularly through pre-established and applied “styles”. Chapter 3: The Book of Worms investigates AI’s relationship with contingency, novelty, and undecidability, examining the generative logics of large language models and, more broadly, stochastic AIs. It critically interrogates the claim that AI is inherently regurgitative, questioning whether synthetic intelligence can produce logical leaps rather than merely recombine learned patterns. Through experiments with Polymorph, a physical, complex adaptive AI system developed collaboratively at AIDLab, the chapter argues that the logic of reassembly is not inevitable. It demonstrates that AIs attuned to their physical circumstances and not constrained within linear architectures can develop emergent, self-organising, and exploratory behaviours.

Funder: London Arts and Humanities Partnership LAHP, AHRC [2243423]