A Case Against AI Film

And misunderstanding the art of film

2001: A Space Odyssey. Source: FILMGRAB

“Anyone can be a filmmaker. This democratizes the craft.”

“We can make sequels to films that don’t have them and more episodes of cancelled TV shows.”

“We can create films based on exactly what we want to see and don’t have to share it. The same film will never be watched twice.”

These are all paraphrased sentiments I’ve seen circulate Twitter/X in the last two weeks regarding OpenAI’s new text-to-video generative software Sora. Mostly by “tech bros”, people who mindlessly glorify AI and its innovations without considering ethical concerns. People who have no interest in understanding why AI should be regulated as soon as possible. People who see a shiny new sharp toy and, without thinking about the repercussions, pick it up and start to play.

This subset of people, pioneering the indulgence in AI we see in online spaces right now, are fundamentally misunderstanding what makes a film. It isn’t a purely aesthetic pursuit, it is hundreds upon hundreds of individuals coming together in an uncertain present and creating a singular vision. It intentionally conveys meaning through every aspect of itself; lighting, blocking, acting, writing, and camera work, to name a few. It’s a vision being brought to life by real human beings, with hands, hearts and brains, working together to create impossible results.

In OpenAI’s announcement thread on Twitter/X, they linked several videos and their prompts. A snowy Tokyo street. A trailer for a film about a spaceman shot on 35mm. A papercraft coral reef, an animated monster, a woman walking at night in Tokyo. These linked videos are conceivably impressive on first viewing, but there’s something sinister about their uncanny valley nature, about the meaningless movements that characters act out, a wretchedness in their lifelessness. Interestingly, a central argument online is that AI allows those without technical ability but with creativity to “break into” film. Yet no human hand has touched these videos except to suggest what it would like to see. No genuine artistic intention lies beneath their surface, much like AI visual art; the characters that populate them are dead upon arrival.

[embed]

“Anyone can be a filmmaker. This democratizes the craft.”

Sora does not democratize the craft of filmmaking. It allows those without true skill or passion to indulge themselves in unartistic, purposeless recreations of real film. The democratization of film already exists, found in the devices we hold in our very hands right now (and before you say it’s impossible, 2015’s critically acclaimed Tangerine was shot with three iPhone 5Ss and made 935,900 USD at the box office). The idea that with the means comes the result is untrue, as to be in any role on a film set you have to work your way through small projects, hone that specific skill, and create meaningful human connections. You can’t buy a boom mic and call yourself a boom operator because you hold it sometimes alone in your room. I’m not saying everyone who shoots a film on their iPhone can and will find success, but my point is it can be, and has been, done within the industry before.

“We can make sequels to films that don’t have them and more episodes of cancelled TV shows.”

Using AI software to create sequels of films or more episodes of shows that have been cancelled or have ended is an empty pursuit. Without the creative minds, the human talent in front of and behind the camera putting their intent into their work, what is the merit of an extra episode of a show? If an extra episode or season of Succession had been beneficial to viewers and served the writers’ visions, it would already exist. It not only ignores but also insults artistic intention.

“We can create films based on exactly what we want to see and don’t have to share it. The same film will never be watched twice.”

Suggesting no one film will be watched twice, as we will be able to “create films that only we want to watch”, extracts the sense of community from film. It ignores the novelty of sharing a packed cinema and laughing and crying alongside other human beings. It disregards the craft of film criticism. There is also a danger to only consuming “exactly what we want to see” — to limiting our consumption of art to our ideals and comfortability. Having something made specifically for you by AI ensures you will not be challenged or offered a divisive opinion when real art should continually be striving to do that.

People who champion Sora as the future of film fail to recognise that art without human intention is not art. Art without shared experience is not art. AI videos will never be able to convey human emotion the way a real actor, in front of a real camera with real people behind it on a real film set can. AI cannot replicate the decisions people who are masters of their craft make on set. It cannot understand the nuance of character, the subtlety of emotion, the intentionality behind light and colour and framing; behind the mise-en-scène. This software can draw from all corners of the internet and swallow down as many screenplays as possible, but it will never replicate real human speech patterns or emotional beats like a screenwriter can. The costuming, the make-up, the set design, the nature of the actor's relationships with each other onscreen — nothing this human can be manufactured by something inhuman without losing its intrinsic value as art. I’m unsure if AI film will ever even get its start (hopefully not), but while the topic is circulating and the tech-bros are getting far too excited, I felt it was important to discuss it. I find it unlikely that entirely AI-generated films will ever take off past a certain point, but it is more likely that below-the-line workers within the industry could see their jobs diminish as AI improves, which is just as detrimental to the art form.

I’ll end on the note that categorising film as something that we mindlessly absorb, “media” we consume, allows people who develop and use AI, who advocate Sora, to believe they can sell us cheap slop as a substitute for authentic art because they see the medium as nothing more than a means of generating money. Only when we accept it as something we can use to purge emotion, to communicate the human desire to be seen and understood, to come together and create a precise image that can convey something different to every person who experiences it — only then can we understand that it can never be replicated by a machine.