📽️ Filmmakers should become centaurs

At Sweden's equivalent of the Academy Awards, the Guldbagge Awards, filmmaker Peter Dalle gave an impassioned speech about the human brain. About human creativity. Unfortunately, he made the mistake of positioning it against AI. As if they were opposites. Instead, it's the contrary. AI unleashes human creativity and potential.

Lingonberry jam?

Dalle presented the award for Best Screenplay and began by talking about lingonberry jam. That's what he thought about when looking out at the audience. His point was that this was something only a human brain could come up with, and no AI. Because the human brain is unique. Only your brain has had your upbringing, your loves and sorrows, and therefore it's the only one that can create something entirely its own.

If you give an AI a hundred episodes of a TV series, it can probably piece together a hundred and first that largely resembles the first hundred, Dalle argued. But do we really long for a "Love is Blind Finland" or "Dancing With the Stars in the Gobi Desert"? he asked.

Peter Dalle is right about the brilliance of the human brain. Dalle has achieved legendary status in Sweden with a long line of funny, strange, and popular creations in film and TV. There's a special brain in that head. But otherwise, there are flaws in his reasoning. Even though I agree with Dalle about human creativity, we see, not least in the film industry, a severe lack of exactly that. Of the 20 most-watched films in the world last year, only two weren't sequels. In Sweden, there was only one such film in the top ten. What Dalle claims AI would produce – more of the same – is exactly what we see from humans.

AI made artists more creative

One way to remedy the lack of new thinking is to use AI in creation. This is where Peter Dalle, and so many others, go wrong. They see it as either/or. AI or human. Human or AI. The answer is both, together. It's when human intelligence mixes with artificial intelligence that magic happens.

This is evident in a research study of four million artworks from 50,000 artists, where researchers compared artists who didn't use AI in their creation with those who did. Those who used AI were generally both more productive and creative. Those who didn't use AI produced fewer and more uniform artworks. The study becomes most interesting when we examine the differences among those who used AI. Some artists became more creative and created more unique works, while others pumped out increasingly uniform and generic art.

The artists who became more creative in their work embraced the technology as a collaborator rather than a tool for mechanical production. By letting AI generate new ideas, they took the first step toward something unexpected, but crucial was how they themselves then shaped, challenged, and refined these suggestions. They only published what had a clear voice and thoughtful expression. The interplay between human and machine gave them access to a larger ideological universe than they could explore on their own. AI became a catalyst, but it was the artist's own thoughts and ambitions that gave the works their depth and meaning. Experimental spirit became a key factor – a willingness to try styles and themes that lay outside the expected, where each AI-generated suggestion was a starting point rather than a finished result.

On the opposite side of the spectrum, we find artists who instead used AI as a convenient solution. Their creation was characterized by a tendency to accept AI's suggestions directly, without critically examining or further developing the material. These artists often stayed within already established aesthetic and conceptual landscapes, resulting in works that reinforced the expected rather than challenged the given. By using the same or similar instructions to AI, they risked getting stuck in a stylistic homogeneity. The focus in many cases was on productivity – creating more works in less time – rather than experimenting or creating something that stood out. Many adapted to current trends, which certainly gave them an audience, but at the cost of the artworks' distinctiveness and originality.

Peter Dalle should become a centaur

That human intelligence together with artificial intelligence creates the best results is visible in several areas. There's a type of chess called centaur chess. There, humans play in teams with computers against each other. The computers on their own are superior to humans in chess, but if humans are good at using their own intelligence together with the chess computer's, they win. These players are centaurs, half human, half computer.

In recent months, generative AI models like Sora and Veo have made it much easier to produce high-quality video based just on a text instruction. It's an excellent tool for screenwriters to quickly and easily test new concepts and ideas. They can even throw in Peter Dalle's lingonberry jam and see what happens.

Mathias Sundin
Angry Optimist