AI and the Survival of Cinema | Evelyn Kreutzer

A community in dialogue

“Future of survival” – a catchy title for a three day event and a conference at the Locarno Film Festival. Almost too catchy to not suspect it as a product of concept marketing: is there a mass-mediatic technology of fear behind the choice of the word “survival”? Is there the shadow of dubious vapourware in any proclamation on “the future of”? I don’t want to be a master of suspicion but prefer checking the facts with candid curiosity. At the Locarno BaseCamp I discover that there is no know-it-all authority telling me “Alarm! Alarm! I’ll tell you how to cope with the danger!”, but rather a passionate analyst of contemporaneity named Kevin B. Lee, who has been able to gather scholars, artists and cinephiles around “urgent” (hence survivalist) themes such as the systematic destruction of social and ecological equilibria affecting the transformation of cinema. With his team at the University of Lugano, they create a welcoming community of people interpreting “the future of”, not in terms of ideological recipes but of imaginative experiments and sincere dialogue.

AI-enthusiasm and “retro-socialism”

One of the three urgent topics raised involved technology, more precisely the use of Artificial Intelligence (AI) in generating film («AI and Generative Humanity»). Is cinema going to survive the new technology? Or could it survive with or even thanks to this technology? Paul Trillo is the expert practitioner whose enthusiastic experiments with AI triggered the discussion in Locarno, and had the participants express their fears or concerns towards the new technology. Beside the typical fear of an out-of-control technology stealing our creativity, and the more general technophobia revealing the naïve and romantic presupposition of free and natural art-making, the most interesting criticism of AI-generation in film (introduced specifically by Richard Misek in the roundtable) tries to dismantle the accessibility-therefore-democratisation argument provided by such technology, and underlines the hi-tech oligarchy of big data companies as well as the power consumption of their engines – not restricted to the alleged obscurity in their management of human labour. All of these preoccupations seem to be well-grounded, yet still lacking comparative studies that demonstrate that a world without AI-technology would allow better situations, both in political power distribution and physical power consumption. Does a Hollywood, Bollywood or even an arthouse film supported with public resources really have a more sustainable social and ecological impact than an AI-generated film?

I was astonished by the easy consensus that the audience shared on conservative measures as limitations, regulations and AI-taxes. I have the impression that many left-wing intellectuals are currently trapped in what I would call “retro-socialism”, dreaming of a regression to more natural styles of life, without questioning the mythical notion of “nature” as a place of harmony. On the contrary, a truly socialist agenda (to my eyes) would need a collective appropriation of the new technologies – learning programmes in public schools for example – and experiments of sharing – not only open source software but also non-privately managed hardware.

Expanding with Evelyn Kreutzer

These impressions on the roundtable that Paul Trillo [stills from his films above] hosted together with Miriam De Rosa and Richard Misek are the starting point of the podcast discussion I would then have with Evelyn Kreutzer, film scholar and video-essayist, working with Kevin B. Lee in the research team at the University of Lugano. Together with the general topics I mentioned here, we expanded on more specific aspects of the aesthetics of AI-generated films, also taking inspiration from the muted screening of Trillo’s videos during the roundtable. For example, the fact that no one in the audience asked about the sound of his videos obliges us to discuss the intriguing separate focus that the current AI-tools seem to create between image and sound.

Her expertise as video-essayist also allowed us to develop the intrinsic relationship of AI-technology with past images and archive footage – a topic raised by Miriam De Rosa in the roundtable. Is there a nostalgic vibe in AI-generated films, or a new experience of remembering and misremembering? And – I propose – should the common occupation with archive footage have us imagine an AI-generation model like a gigantic video-essay machine – if a machine can ever create essays?

Final plea for medium unconsciousness

The first phase of introduction of any new technology pushes us to reflect on the limitations of the tools, showing them, making them the occasion to explore new possibilities of creativity; this is Trillo’s explicit position and, in this way, the tool automatically becomes a medium, possibly a self-reflective one. Nevertheless , the rapid development of AI should remind us that a medium is always also a tool, and that reflection (on the medium) cannot prevent appropriation (of the tool), that a theoretic approach cannot pre-empt the aesthetic practices seeking to embody visions. How not to be ignorant of the medium but still have the possibility of experiencing a moment of medium unconsciousness: this is probably the future of cinema, surviving technology by appropriating it, this is the future survival of cinema through technology.