VERDICT: In partnering with Google’s Image Recognition AI, Jeppe Lange has constructed a 100mph frenzy of match-cutting that is strange, rhythmic and at times somewhat profound.
Over the past few years, more and more filmmakers have been exploring the potential ways that artificial intelligence can interact with storytelling. In films like Chris Peters’ 24,483 Dreams of Death or Amy Cutler’s All Her Beautiful Green Remains in Tears, algorithms are fed data and charged with conjuring new forms of narrative. In Abyss, the new film from Jeppe Lange made using Google’s Image Recognition AI, the ask is perhaps slightly more straightforward, but the results are no less engrossing or profound. Lange’s film consists of a chain of 10,000 images selected on the basis that each image visually resembles the one preceding it.
This may sound like an optical assault – with the film lasting just over 13 minutes it works out at more than 12 images per second – but the eyes and mind quickly adjust. In the same way that our mind fills in the blanks between the static individual film frames to simulate cinematic movement, so here our mind registers the patterns that emerge in the gaps between the AI’s selections. Some of the transitions are fleeting and feel like fun visual gags – the pattern on a peacock’s tail feather becomes a top-down view of an ominous digital weather front while a series of turtles in the ocean is interrupted momentarily by a pineapple floating in a swimming pool – while others seem to feel more politically loaded or artistically familiar.
When the sharp lines of a series of tower blocks morph, via the tessellating patterns of glass domes, into a chain-link fence, our mind can’t help but search (or perhaps stretch) for meaning in the correlation. When a close-up of a vagina cuts to a blossoming flower, a centuries-old art history metaphor is accidentally referenced by an algorithm unaware of its prevalence. Filmmaker Jeppe Lange describes the ‘misunderstandings’ inherent in the AI’s reading of pictorial media, but somehow this makes the montage feel all the more human. In Abyss‘ coldly analytical trawl for patterns, each audience member projects their own meaning onto its hypnotic, flickering avalanche of imagery.
Directors: Google’s Image Recognition AI, Jeppe Lange
Producer: Jeppe Lange
Sound: Simon Brinck
Venue: CPH: DOX (NEW: VISION)
No dialogue
13 minutes
