How ChatGPT Is Undermining the Creativity of Studio Ghibli

Nothing in the realm of cinema is as time-consuming as classic animated films—whether hand-drawn or stop-motion. Both forms require the artist’s physical interaction with the medium, whether it’s clay, a wooden puppet, or a blank sheet of paper awaiting the first stroke of a pencil. Ask Nick Park how long it took him to make the latest Wallace and Gromit movie (six years). Ask Hayao Miyazaki how long he worked on The Boy and the Heron (seven years) or any other animation bearing the Studio Ghibli logo. Patience and meticulousness, the careful attention to each frame—this is the essence of this kind of cinema. If one wants to achieve similar artistic effects, there’s no shortcut: you have to roll up your sleeves and put in the work—or ask the latest version of ChatGPT to generate the desired image.

In the past few days, thousands of people have taken advantage of this possibility. The internet has been flooded with edits that closely mimic Studio Ghibli’s distinctive style. Everything was fair game—memes, photos, favorite movie frames. No hesitation, no scruples. Even Sam Altman, the CEO of OpenAI (ChatGPT’s parent company), openly encouraged this by changing his profile picture on X to an AI-generated image “in the style” of Ghibli. But after a brief moment of collective euphoria, a bitter realization set in, sparked by a handful of skeptical critics: is no one actually losing out in this whole Ghibli craze?

The answer is obvious: as always, the ones losing out are the artists. The people whose hard work was exploited to create these charming yet soulless images. Most of you probably understand how ChatGPT generates images, but it’s worth repeating: artificial intelligence doesn’t create anything from scratch. It pulls from a vast database of images, film frames, and other visual media—most of which ended up there without the artists’ consent—and then spits out a finished product, more or less aligned with our expectations. If the database consisted only of visuals whose authors had explicitly agreed to their use, this wouldn’t be an issue—or rather, it would be a different kind of issue, one centered solely on the question of whether AI-generated “art” has any artistic value. “Art” that Hayao Miyazaki—let’s not mince words—openly despises.

Miyazaki doesn’t hold back when it comes to artificial intelligence. When he was shown an AI-generated animation in 2016, he called it an “insult to life itself.” It’s not hard to imagine what he thinks about AI-generated images imitating his and his colleagues’ work. A unique style, honed over decades, has now been reduced to a few lines of algorithm, fattened on the unauthorized use of Ghibli artists’ work.

Yet it’s easy to downplay the issue, as Polish movie critic Michał Oleszczyk did in a particularly thoughtless post, dismissing it as just another passing trend—arguing that in the end, Ghibli will benefit because people might discover its films, “perhaps meeting Totoro for the first time.” And anyway, there’s no point in nitpicking because “the whole sandbox is problematic, not just this one toy car.” Oh, sweet naivety! The last thing Ghibli needs is free advertising built on the theft of its intellectual property. Especially considering the studio’s deeply ingrained distrust of technology, which has led it to minimize its use to the absolute essentials. Except for the disastrous experiment that was Earwig and the Witch, directed by Miyazaki’s son, every Ghibli film frame is hand-drawn before undergoing digital processing. Generating images in their style—whether we like it or not—is a complete contradiction of the studio’s values and the philosophy of its artists.

The outrage expressed by Ghibli fans worldwide is justified. That one toy car—Oleszczyk’s metaphor for this whole affair—has, in fact, exposed a broader ethical problem regarding the use of artificial intelligence. A problem rooted in the gray areas where these AI tools currently operate, particularly concerning copyright and environmental impact (as a reminder: generating a single AI image consumes about several dozen milliliters of water for server cooling). Those benefiting from these unresolved issues are everyone—except for our dying planet and the original artists and writers. Instead of lamenting the selective outrage of internet users, it’s better to appreciate the slight increase in public awareness and engage in a discussion whose outcome could shape the future of AI. After all, cleaning up a sandbox always starts with picking up the first toy car.

And one last thing, a bit of an aside. For centuries, historians, philosophers, and other scholars have debated the role of inventions in human life. As Nicholas Carr writes in The Shallows, a book analyzing how the internet affects human cognition, these debates tend to divide people into two camps: instrumentalists and technological determinists. The first group asks, What can humans do with technology? They see innovations like the internet or AI as mere tools, completely subject to human will. As Carr notes, “Instrumentalism is the most common way of thinking about technology—especially since it’s the view we would prefer to be true.”

Personally, I find the deterministic perspective far more interesting. In this view, the question is flipped—we no longer ask what we can do with technology, but what technology does to us. How do specific inventions shape our cognitive processes, our creativity, our intelligence? What are the long-term costs of using them? It’s worth considering this the next time we ask ChatGPT some trivial question or request an illusion of engaging with the art of our favorite creators. Because who knows? Maybe Miyazaki was right when, after his fateful encounter with AI, he said: “We are nearing the end of our times. Humanity has lost faith in itself.”