Weaving Worlds: how artificial intelligence can push human creativity

Weaving Worlds is a collaboration between Amanda Baum, Rose Leahy and Rob Walker and is an exploration into how nonhuman logic can act as a catalyst for collective making and thinking. The publication is split into three individually written texts which underwent a process the trio named “weaving,” to produce a publication that demonstrates how humans can push their creativity by incorporating a nonhuman element.

After every iteration of the publication, a machine learning algorithm called word2vec, analysed the texts and pulled out latent relationships in the language, turning these into a list as well as a graphic visualisation. These relationships appear random and unconnected to humans as they require massive amounts of repetitive statistical analysis and a perspective that’s not possible by human processing of language. The results of this “weaving,” both informed and inspired the content of the publication.

Weaving Worlds is essentially an experiment looking at what happens when singular human authorship and creativity is disrupted by artificial intelligence (AI) and how this can push the boundaries of what we as humans have the potential to create. “We were each interested in what happens to both our individual and collective creativity when agency is given to the nonhuman, whether that be an algorithm, another species, a chemical reaction,” they explain. While writing essays around this concept, they wanted to find a method of making that mimicked the topics they were discussing, allowing the project to become an example in itself of what they were advocating.

They decided on a process that would reflect their concept of collective making in all areas. While they, of course, drew on their areas of expertise – Amanda in design, Rose in fine art and Rob in technology – all parts of the process went through the same process of weaving to create an outcome that represents a true collaboration. “The idea emerged from a shared wish to evolve the typical structure of an individually authored dissertation in order to embody the discourses that underpins our work,” they explain when asked about this unique way of working.

The design of the publication is largely dictated by the visualisations created by the AI, allowing the book’s aesthetic to mirror themes of weavings, meshes, webs, and networks that occur in the text. “In a similar way that we wanted the entire structure of the book to reflect the content, the graphics grew out of the actual visuals produced from the technical processes.”

The project began as the trio’s MA dissertation while studying at the Royal College of Art but they have since developed the concept into a framework for collective research, which they’ve tested in workshops, seminars and panels. They are currently in the early stages of publication with Danish publishing and curation platform Laboratory for Aesthetics and Ecology.