How would I walk, had I never seen a woman walk?

Installation, Walking Pad, Camera, Video: 6:35min

Every person has an individual gait which serves as a biometric feature for gait recognition. The result can be supplemented with the system of automated gender recognition (AGR) classifying people into the binary system of man/woman. These processes are based on datasets – collections of annotated image/video data used to train artificial intelligence (AI) systems.

The video work criticizes the automated binary gender attribution and its consequent misinterpretations. It problematizes that "gait recognition" combined with "automated gender recognition" leads to an invasive surveillance method with datasets never being fully diverse. How would I walk, had I never seen a woman walk? contrasts the strict AI systems with the fluidity of gender.

In installation, the work consists of found footage from scientific gait datasets, screen recordings and new videos, combined with spoken dataset information and music (by atorbit). Exhibition visitors trigger both the work and a video recording by walking on a treadmill. These videos form a new dataset of surveillance videos that can be seen on the back wall.