Director: Richard Levine
Running time: 1min
With James Cameron’s latest box office behemoth still raking in cash around the world, I have been thinking about the original Avatar a lot in recent weeks. In particular, I find it fascinating that in the wake of the film’s success, there was a glut of 3D cinema – and that nothing came close to capturing the imagination of audiences.
It occurs to me that this is due to two interlocked factors. Studios, in their haste to pump out products which could cash in on what they saw as a new trend, underestimated Cameron’s ability to use technology to tell a story that impacted people in a way other films had not – and overestimated the pull of the technology itself. No amount of “it’s coming right at me” trickery could make more people want to see Piranha 3D, or the many, many other blatant cash-ins that accompanied it to theatres.
This is something which filmmakers might want to bear in mind when it comes to the implementation of AI. The technology enthusiasts rushing AI-generated imagery to theatres as stand-alone films may be blinkered to the fact that while they might find the new mode of working very interesting, when the initial novelty wears off, audiences probably won’t.
The Nun is the second AI-generated film received by Indy Film Library – and the first to have actually met our criteria when it comes to clarifying who created the art it features. Meeting our requirements that the imagery and sound were obtained consensually, director Richard Levine confirmed that he created all the initial visuals with the exception of one analogue image that was a public domain picture of a generic “1920’s European nun”; while the sound track is an unaltered complete piece by drummer Josh Freese, used with his permission. For that reason, The Nun will be receiving a score – making it by default our most positively-reviewed AI film to date.
It’s not just by default, either. Freese’s soundtrack in particular is a delirious joy to hear; an absurdist free-form jazz piece, complete with a bizarre, honking riff on The Glenn Miller Orchestra’s In the Mood. At the same time, the visuals themselves are vivid, lively and at times unnerving. The imagery deployed by Levine is at times disarmingly childish – a mash of interwoven splodges and stains – but at others, it features the mad, staring eyes and withered faces of a David Firth animation.
It is a discombobulating juxtaposition, that grows into a unique and eerie synthesis after being fed to the Midjourney AI. Animating the images together into a fluctuating mass, there is something almost Cronenbergian about the end result – with pieces of innocence absorbed into the malevolent flesh of some passing monstrosity. On that basis, I have to admit that I kind of enjoyed it.
The problem is, that kind of hinges on the fact that at no moment did I ever feel like I had not seen this kind of film before, or that I hadn’t seen it done better. The endless evolutionary odyssey of Joanna Priestly’s North of Blue delivered so many remarkable and impossible images, which were blended seamlessly together into a single, teeming ecosystem – and that all happened via her own work as an animator. Meanwhile, Marijke de Belie’s Feminam showed us defined and recognisable iconography around women throughout the ages – including the image of a nun – which waxed and waned to deliver a thought-provoking, hand-drawn, experimental film, wrestling with identifiable elements of the human condition.
The Nun does flow in a way that those do, using animation to assemble aesthetically jarring imagery into a singular whole – but it does not seem to be to any thematic end beyond seeing if it were possible for technology to do this. According to Levine, his key goal for this project was to see if he could “manipulate Midjourney to work in an abstract, non-linear, non-narrative way”, contrary to its programming, which primes it to “create understandable narratives” or “replicate a literal world”.
Of course, there is no objective standard with which to determine if that goal is worthwhile pursuing. To some people, it will be, and I respect their desire to explore the limits of AI’s imagination, or its capacity to create abstract communications. But it doesn’t especially grab my interest at the moment, because in contrast to the other abstract films mentioned, it will not come from a thinking, feeling being (yet). There is nothing to empathise with, to try and understand what needs the creator of AI art might be trying to serve, beyond the commands of its masters.
I am not saying that I do not foresee a time in which AI art could fulfil those criteria, or that I do not believe the technology has exciting implications for independent filmmakers. It’s just that here, the process seems to be the most interesting thing on offer – while the product it has yielded is not especially exciting.
This is not to say that Levine or other artists are barking up the wrong tree by seeing what AI can do in terms of experimental film. But a longer form (one minute is not long enough for us to really immerse ourselves in whatever imagery is provided) and trying to find a way to allow the AI itself to be a bit more introspective – if that is even possible – might take this format in a more engaging direction.