Mapping Latent Reality In Search of Forward Remembrance

Abstract

In conceptualizing an approach to the use of generative AI that is steeped in an artistic research paradigm, we introduce Forward Remembrance, as praxis for navigating the space between what AI models have learned and what futures they can imagine. Using concepts evolved while building Synesyn, a practical suite of tools, we demonstrate how to move through this "latent reality" - the encoded potentiality inside generative models - as artistic research. We walk through a conceptual and demonstrative outline of the underlying ideas, showing how 'navigational acts' can help steer the process of inference beyond where the AI's default ‘closures’ are wont to lead. The contribution is methodological and is aimed at turning speculation into inference, with the AI models as tool and the artist as navigator.