Music | This is what a music video created with artificial intelligence looks like, according to director Juho Lähdesmäki, artificial intelligence “changes everything in the creative field”

Magical realism, David Hockney, jasmine and cherry blossoms, night, oil on canvas. Among other things, such prompts or commands have become one of the first known music videos created with artificial intelligence.

Of course, making a music video required a lot more than just inputting commands to artificial intelligence. Emma-awarded director Juho Lähdesmäki made precise determinations for each frame about the movements of the “camera” and how much artificial intelligence changes the next image, what kind of subjects, colors, lights, shadows and sharpness are visible in the images.

The whole concept reflects the world that Lähdesmäki and the band wanted to see in the video. “Artificial intelligence has no imagination,” says the director.

Juho Paalosmaan and Jori Hulkkonen synthpop duo Sin Cos Tan for the chapel Endless the music video made was released this week.

Watch the video below:

Lähdesmäki is on the crest of a new technological wave with his video. Highly developed and efficient text-to-image artificial intelligence codes and generators that utilize them have started to be released for free use within just a few months.

In the past, the mechanics of artificial intelligence art have been under the commercial control of a few technology companies such as OpenAI and Google. At the end of August, the Stable Diffusion code was announced for open use, which Lähdesmäki also used to create the music video.

Working on the video started with an intensive learning of artificial intelligence tools. The director also discussed the themes of the upcoming album with the band. Eventually, he began to gather ideas around “middle-class dissonance”.

From the sliding image stream, you can distinguish the big city milieu, leisure time at the swimming pool, the beach and the golf course. Cracks appear among the image of pastel shades and abundance. Lähdemäki reveals that these pictures were created by the order: “Garden party during a riot”.

Lähdesmäki fed commands to the artificial intelligence, based on which it borrowed style, subjects and forms from, for example, famous artists such as David Hockney, Nicholas Roerich, Rob Gonsalves, Tim Hildebrandt. The final result does not show a single copy of the artists’ works, but traces of, for example, the formal language of pop art and modern expressionism or the way Hockney describes swimming pools can be seen.

Artificial intelligence art tools have not yet been created for a moving image, but Lähdesmäki created a series of images with additional code that simulates the movement of the camera and enables iteration of the image in relation to the previous one.

He calls the end result the machine’s fever sleep.

“When I went to sleep, I set the machine to render the video for the night, and I had no idea what kind of dreams the artificial intelligence could have seen during the night,” he describes.

That an image can be created on command, has already in a short time revolutionized the way of thinking about the visual environment and the work of the creative industry, from graphic design to games, films and visual arts. It has already been used in making films, on magazine covers and in advertising – and now in music videos.

“This will change everything in the creative field. The code can be used in many and surprising places,” says Lähdesmäki.

Artificial intelligence tools have also caused concern and anger.

For example, an image created with the Midjourney program won recently art competition in Colorado, which sparked anger among artists and an existential debate about what artistry means. Our entire concept of art is based on human creative activity. Is a person feeding commands to artificial intelligence an artist? Or is the artist in this case an artificial intelligence?

In the creative field, there is a fear that a technology like Stable Diffusion will take work away from people, while it uses copyrighted footage, an image made with the work and effort of professionals. Magazines dealing with technology culture, for example, write about the topic Wired and The Verge.

Lähdesmäki was also worried at first.

“The further I got and the more I learned, the more I understood how much that tool can be controlled in the desired direction. At first, the images became quite random. It changed my own attitude. This is just one more tool,” he describes.

Artificial intelligence ultimately does not understand what is good for its user.

“We need an artist to mirror the world we want to create.”