AI as spatial practice

Almost 30 years ago architect Jane Rendell characterised critical spatial practice as site-specific acts of reading and writing situated between art and architecture. Around the same time, AI researcher Philip Agre called for a critical technical practice: ‘Technology at present is covert philosophy; the point is to make it overtly philosophical.’ Bringing these together, we investigate how a critical spatio-technical practice might effect public exchange and opinion formation about urban spaces, in order to enable more empathetic and inclusive ways of living together.

We explore this through practice, harnessing AI as a technical tool and partner in helping interpret and reimagine urban public spaces through critical site reading and writing, based on philosophical concepts. This extends work undertaken during a SPACEX secondment, and aims to reconfigure machine learning and computer vision for interpretation, not classification, engaging with the ambiguity and cultural specificity of real human experience.

We do this specifically through a focus on performativity in moving image, as both input and output, training an AI model not to classify specific objects or actions in urban spaces, but philosophical concepts and affective relations. We then reflect the outputs to urban inhabitants to prompt public exchange of opinions, aimed at empathy and inclusion. By foregrounding philosophy and challenging AI to embrace ambiguity and complexity, our approach aims to co-opt and reposition AI as a partner in the interpretation of real human experience.