Fields of View contines research into the relationship between an algorithmic acceptance within society and the perceived acceptance of imagery when presented in emotive material forms such as 35mm film. Film is a connection to lived experiences that we can subjectively relate to in the pastoral nature of sky and mountains. Whereas it is in tension with the more clinical nature of data and our acknowledgement of place as a political entity.

The intrinsic, tangible quality of celluloid film and paper print play to our sentimental memory bias making it harder to distinguish between what is real or fake. We as humans take on the role of the neural network (“brain”) predicting the extent of ambiguity. Using new creative methods of transforming neural networks, the process analyses not only the generated imagery but also manipulates the features in the neural net, isolating specific ‘known’ aspects of an image dataset (a method known as network bending [1] ). For example, how would it change our interpretation to isolate "semantic groupings"[1] that include only the sky or only the mountains of a specific narrative?

By transforming certain features in the data the work makes comparison to historic in-camera editing techniques of cameraless film making.