CODIFYING CLASS 

Exhibited at South London Gallery.

7min video + interactive code

 

This research project was made during a 3month residency with school of speculation in partnership with The Design Museum, Koppell Project, Bold Tendencies Institute of Contemporary Arts (ICA and South London Gallery, 

​

'Codifying Class' harnesses classification and categorisation algorithms trained on labelled imagery, and highlights the hazardous implications of the bias simplification of groups in society. Utilising tools such as scraping Google images and computer vision, they can provide an artistic insight into how computer systems are being trained to view marginalised groups while adding to the wider discussion of who is to be held responsible when these technologies don’t serve the community. 

 

The interactive piece scans the viewer's face and attempts to label you as working-class or upper-class. I trained the algorithm making these predictions on a dataset made of scraped images from Google that were labelled working-class. Obviously, there is no way to ‘tell’ if someone is working-class. However, this interaction prompts the viewer to question; how they feel when prescribed a label? If it gets it ‘wrong’, will it make you question the way you’re perceived? While highlighting the possibility for the input of human bias into technology. 

​

The video loops through the database created from the scraped google images labelled working-class, providing insight into what the algorithm is trained on and inviting the viewer to explore the databases imagery,

​