In an era where AI reshapes every facet of society, polarized public debates often frame AI as either an impending superintelligence, akin to Hobbes’s Leviathan, or as narrow AI covertly manipulating our digital lives. This thesis challenges this dichotomy by investigating a third scenario: ‘being-with’ algorithms in artistic practices. The research explores how collaboration between sentient beings and algorithmic intelligences can result in dynamic, creative assemblages characterized by complex, non-linear behaviour — a phenomenon identified in systems theory as a complex adaptive system.
The research adopts an interdisciplinary approach, drawing on philosophical concepts such as Duchamp’s found object, Golding’s ana-materialism, and Deleuze’s notion of difference, alongside scientific frameworks like Downing’s investigations of emergent phenomena through complex adaptive systems. Central to this inquiry is the development of Libidinal Geometry, a novel framework that combines Mandelbrot’s fractal geometry with Lyotard’s libidinal economy. This framework bridges libidinal, discursive, and mathematical expressions, adapting Hayles’s non-conscious cognition into non-sensuous expression. The research redefines the role of machines from mere tools to co-creators, employing deep reinforcement learning agents that navigate complex sentient environments and interpret biometric data such as excitement and fear.
This exploration culminates in the project Abysslips, where frameworks of algorithmic and sentient libidinal forces are interconnected within a computational sensuous environment. Biometric sensors capture human affective states, forming mathematical tensors that agents of advanced reinforcement learning explore through deep neural networks driven by self-play and intrinsic curiosity. These agents also analyse the simulated physical environment of 3D printing, learning the grammar of G-code to infuse figural qualities into the discursive flow of the printing process. By translating four-dimensional experiential flows into three-dimensional expressions, the resulting sculptures emerge as material articulations of complex, multidimensional libidinal processes rather than mere replicas of idealized states. Although the project’s full realization extends beyond this research’s scope, it establishes the necessary algorithmic, philosophical, and material conditions for future implementation.
The outcome of this investigation reconfigures how art is conceptualized and created in the digital age, presenting a new mode of emergent art that predicts significant shifts in future artistic practices. Future work will continue to explore the dynamic interplay between sentient beings and algorithms, pushing the boundaries of artistic and technological collaboration.
Keywords: algorithmic expression, 3D printing, complex adaptive systems, emergent phenomena, biometric signals