Knitting Algorithmic Assemblages

Abstract

This paper aims to explore knitting as a metaphor for raveling and unraveling the emerging relationships within assemblages of human and machine intelligences. Centered on a practice- based investigation into the generation of knitting machine punchcards through the use of machine learning models, the project reveals the co-creative capacities and implications of such assemblages. Knitting is a process of creation, where interlocking loops of yarn repeat to generate a surface. Repetition through patterning produces complexity and texture; knitting itself is an assemblage including human, technology and surface. Artificial intelligence (AI) and machine learning (ML) are human inventions around which orbit various metaphors of acceleration, uncontrolled emergence and ecology. These metaphors delimit our understanding of machine intelligence, positioning it as separated and extractive. In this paper, knit is employed as a metaphor and a practice to locate the looped and interlocking generative nature of human and machine intelligences. This metaphor also joins computing to its shared history with textiles through punch cards and the Jacquard head. The knits in this project materialize a relation between machine and human intelligences. Human subjectivities manifest within their patterning, questioning the perceived impartiality of artificial intelligence that is often bound to contemporary metaphors.