Historically, urban environments have intentionally and unintentionally provided social and public spaces. As a pervasive surveillance state encroaches upon these spaces, it is harder to highlight and identify those areas that are actually ‘public’. This project takes the Situationist notion of a dérive and recontextualizes it for an age where computers can dream. The contemporary city is continually under surveillance, and public spaces have been radically transformed and mutated. As part of this transformation, intelligent machines have extended their reach to the physical world. Facilitated by a global network, but also cordoned off by private entities, the data aggregated through networked objects and locative media has created a new hybridized space where people can be tracked from afar as they move through an urban landscape that is both public and private. Using security camera footage from insecure IP cameras around the world as training data, a computer creates a surveillance map of a city imagining what these public spaces appear like to the private organizations and citizens who have access to the ‘real’ feeds of images.