All Square

All Square is an ongoing research project that deals with humanoid agents, that learn to move independently in various 3D environments using machine learning algorithms. In various training areas, AI agents are systematically trained to learn everyday tasks such as navigating through rooms. The spatial scenarios vary between reconstructed industrial halls - referenced by artifacts and material structures - and highly abstracted 3D scans, reminiscent of captchas, whose landmarks are fragmented. Some levels stretch into infinity, while others remain limited to a few square meters. Theoretically, All Square ties in with Jean Baudrillard's concept of simulacrum and hyperreality. While the industrial sceneries can be regarded as alienated images of a physical reality, the abstract levels operate beyond any concrete reference and mark a phase in which simulation and symbolic worlds have taken over the role of the “real”.
“All Square” reflects on current developments and applications of AI technologies, questions our relationship to digital twins and examines the aesthetics of machine learning.

The site specific video installation for the Luminafest Amberg is set in a virtual factory that creates digital twins. In various training areas, AI agents are systematically trained to learn everyday tasks such as navigating through rooms. Initially, their behavior appears clumsy and artificial, but after a certain period of time, human traits become recognizable in the processes and autonomous movement sequences emerge. The constantly changing factory environment makes it difficult to recognize and classify the agents' movements. “All Square” reflects on current developments and applications of AI technologies and questions our relationship to digital twins as well as virtualized and automated work.

Produced for the international light art festival Luminafest 2024 in Amberg.