Paper short abstract:

In this paper we discuss how Surrealist techniques of automatism, geared toward unlocking the unconscious, can be combined with the supervised machine learning of Artificial Intelligence, to excavate and generate 'creative accidents' within a collection of archival, amateur, found-film footage.

Paper long abstract:

In her classic text 'Melancholy Objects' (1973), Susan Sontag describes how the French surrealists viewed photographic technology as 'liberating' because of the automated machine's unique capacity to manufacture images transcending mere personal expression. Pre-modern visual technology, Sontag writes, is particularly interesting, because its automatic, electronic qualities provide fertile ground for the unexpected and unintentional 'creative accident.' No matter how much control a photographer possesses over the machine, certain elements of the image-making process invariably slip through their fingers; the optical/chemical/electronic impulse gives rise to a new relationship between images and reality.

Our presentation takes as its starting point a collection of found, amateur, archival film footage originally belonging to a family of Eastern-European Jewish immigrants to New York in the 1920s and '30s. In our collaborative work of revisiting and re-editing this material into a new film, we draw upon Surrealist ideas of the 'quasi-magical, quasi-accidental' qualities of photographic images that are mediated by both human and automated forces to produce meaning and memory.

Experimenting with integrating new techniques of Artificial Intelligence into our editing process, we are interested in excavating and perpetuating 'creative accidents' within this footage. As the Surrealists sought to examine the social and psychological effects of mechanical and commodified orders (Foster 1991), we are mobilizing contemporary forms of machinic sensitivity to explore the significance of this archival material, not for its specific historical or personal narratives, but for the light it can potentially shed on the haptic, sensorial fragility of memory itself.