Focus: Another Intelligence

Text: Jonas Holmberg

“Now I have become Death, the destroyer of worlds.” The quote from Vishnu in the Hindu epic “Bhagavad Gita” is mentioned twice in Oppenheimer, Christopher Nolan’s blockbuster about the inventor of the atomic bomb. The second time is when J. Robert Oppenheimer witnesses the detonation during the Trinity test, the world’s first nuclear weapon test, and realizes the magnitude of what he has done. He had long been worried about where his research would lead, but he continued his work anyway.

Many tech workers in Silicon Valley seem to share a similar feeling. In this year’s Nostradamus report “Everything Changing All At Once”, Johanna Koljonen referenced a study showing that half of those developing artificial intelligence estimate that the risk of their work leading to human extinction is at least ten percent. But they continue their work anyway.

The heavy Oppenheimer mood over the AI question intensified further this spring when the world’s leading AI executives and experts launched appeal after appeal with the message that extinction due to AI must be seen as a threat equal to pandemics or nuclear war, and that development must be paused immediately. Tech companies themselves cannot slow down because we have not constructed a society where pauses or reflection are possible. So everyone continues their work anyway.

What is everyone so worried about then? Just ask a cinephile. We know the answer because the risks with artificial intelligences and thinking machines have been a cinematic constant for more than a hundred years, from silent films like The Mechanical Man (1921) and Metropolis (1927), through debated works like Solaris (1972), The Stepford Wives (1975), Ghost in the Shell (1995), and Ex Machina (2014), to the story of the powerful “entity” in the latest Mission: Impossible film. Our immediate understanding of the AI threat is deeply shaped by cinematic dystopias that have had enormous impact: Blade Runner (1982), Terminator (1984), and The Matrix (1999) all depict future scenarios where the conflict between human and AI forces the heroes to fight against technology and ponder what the real difference between humans and machines is. The most central AI moment in film history is probably still found in Stanley Kubrick’s 2001: A Space Odyssey from 1968, where the spaceship computer HAL 9000 suddenly stops obeying the human astronauts. The line “I’m sorry, Dave. I’m afraid I can’t do that.” marks the moment with machine-like coldness when technology turns against us. This is precisely the moment that many of today’s AI experts fear.

That film history is so packed with AI fantasies, long before the technology actually existed, is probably dependent on the fact that film itself highlights and changes the relationship between human and machine. One can even understand film as a kind of artificial intelligence that independently formulates thoughts about the world. It may sound strange that a film can actually “think”, but given the accelerating AI development, it has probably become more natural to grasp that humans do not have a monopoly on thinking. That there is another intelligence. In any case, one can trace a clear thread of cinematic thinking running through the history of film ideas. It starts with the Soviet filmmaker Sergei Eisenstein’s ideas about how the forms of silent film correspond to the functions of the human brain and continues to the notion of film as a “robot brain” that can create independent cinematographic thoughts, as expressed by Jean Epstein in the book “The Intelligence of a Machine” (1946). As a kind of extension of Epstein’s thoughts, Gilles Deleuze formulated influential arguments in the 80s about film as a “thinking machine” that independently formulates cinematic ideas about time, movement, and consciousness that go beyond what can be expressed with language.

Right now, the relationship between cinematic, human, and artificial intelligence is being renegotiated. This is not only happening in film philosophical seminars but with at least as high intensity within the framework of the labor conflict between film studios and screenwriters and actors that has paralyzed Hollywood for months. The strike began around the same time as tech bosses and professors warned of an AI apocalypse and is largely about how generative AI should be used in film production, and what rights and compensations linked to AI should look like. Should an AI be allowed to trawl through Billy Wilder’s and Nora Ephron’s catalogues and then churn out romantic comedy scripts in a continuous stream? And should an AI-generated Marilyn Monroe be allowed to play the lead role opposite Ryan Gosling? Who should be paid the most in that case?

If this becomes the future, and there is much to suggest that AI will transform almost every corner of the film field, one cannot limit the issues to just equitable compensation models. The advent of generative AI in the production of moving images also raises questions about film as art. Over a hundred years ago, film was accepted as an art form because it emphasized the human behind the machine, that a film is not just a mechanical representation of reality but an expression of artistic processing. What happens when you remove or minimize human influence over the creative process?

Right now, the Danish company Tambo is working on a Werner Herzog film. But it is not directed by a flesh-and-blood Herzog but by “Kaspar”, an AI that has been fed all of Herzog’s works and sources of inspiration. The film team then hands over all creative decisions to the AI, which, of course, has also written the film’s script. Is this art? And if so, is it art in the same way as if the human Herzog had been walking around sweating during the filming? We have long understood art as man’s attempt to understand himself and his place in the world, but such a description is at risk of being outdated as hopelessly romantic.

Just like directing and writing, acting must be understood in a new way in an AI world where it is possible to create digital masks that can replace and merge with any filmed faces. Who is it really that is acting, and how is the relationship between the actor, the character, and the audience who observes and interprets changing? Is it possible to make an authentic portrayal by pasting one actor’s face onto another?

These are some of the main questions in the film experiment Another Persona, an AI-generated version of Ingmar Bergman’s Persona produced by Göteborg Film Festival in collaboration with SF Studios, Gothenburg Film Studios, and the Ingmar Bergman Foundation, which will only be shown once during the film festival. In the work, we use AI technology to replace Liv Ullmann’s interpretation with Alma Pöysti’s new role interpretation of Elisabet Vogler, the mute actress who has lost faith in art and acting and seemingly merges with her nurse.

Another Persona is also the conceptual centerpiece in the section Focus: Another Intelligence, which will also include several new feature films related to the theme, a retrospective of historical AI depictions in film, a program with AI-generated films, and an extensive conversation program where we ask questions about film art, technology, and whether it will be the atomic bomb or artificial intelligence that gets us first.