ChatGPT
Perplexity
Claude
Gemini
In March 2025, OpenAI released an image generation update that let users transform any photograph into something that looked like it came directly from a Studio Ghibli film. Millions of people did exactly that. Social feeds filled with Ghibli-ified selfies, pets, and family photos, all rendered in that warm, hand-drawn aesthetic that generations of people grew up watching. For most users, it felt like a tribute. They were fans doing what fans do: finding a way to inhabit something they love.
Hayao Miyazaki, the co-founder of Studio Ghibli, had already made his position clear. In a 2016 documentary, after watching an AI-generated animation, he described the experience as "an insult to life itself" and said he was "utterly disgusted." He wasn't talking about this specific trend, which hadn't happened yet, but his words arrived in 2025 as if written for it. The gap between how the users felt and how Miyazaki felt was not a misunderstanding, but two genuinely different ways of thinking about what art is and who it belongs to, colliding in public.
That collision is worth sitting with, because it captures something real about the moment creative culture is in right now. The question of what it means to be inspired by someone's style has always existed, but it has never been quite so visible, quite so frictionless, or quite so contested as it is today.
Artists have always learned by absorbing other artists. You study the people you admire, you practice in their manner, you internalize their approach until it becomes part of your own vocabulary, and you develop your own point of view from there. This is how visual traditions are transmitted, and it's how every significant movement in art history built on what came before it. The Impressionists were shaped by Constable and Turner. Warhol was shaped by commercial illustration. The photographers who defined editorial photography in the twentieth century were shaped by painters who died centuries earlier. Influence is not a failure of originality, but the mechanism through which originality develops.
Style itself has never been copyrightable, and that's not an accident. As legal precedent has long held, what copyright protects is a specific expression, not an aesthetic approach. You cannot own "warm, hand-drawn animation with particular attention to the texture of natural light." You can own a specific film, a specific frame, a specific sequence of images. The distinction matters, because it's what has always allowed one generation of artists to build on the work of the previous one without requiring permission at every turn.
But something does feel different now, and that's worth being honest about. When a human artist spends years studying Ghibli films, developing their craft, working to understand what makes that aesthetic function, and eventually producing work that carries traces of that influence, the learning itself is visible in the output. The gap between the influence and the result is filled with the artist's own experience, their own choices, their own failures and corrections. When an AI model replicates a style in seconds from a single prompt, none of that process exists. The aesthetic is reproduced without the labor that produced it, and without any particular relationship to the artist whose work the model trained on.
Whether that difference is a legal problem is genuinely unsettled at the moment, but whether it's an ethical one is a question each creator has to answer for themselves. The practical legal reality right now, as confirmed by the DC Circuit's March 2025 ruling in Thaler v. Perlmutter and the Supreme Court's subsequent decision to let that ruling stand in March 2026, is that copyright protects specific works, not styles. Generating an image that evokes Ghibli is not the same as copying a Ghibli film. But legal permissibility and ethical consideration are not the same conversation, and conflating them tends to flatten something that deserves more texture.
The artists who are most troubled by AI style replication are not, for the most part, worried about legal violations. They're worried about something harder to articulate: the sense that a style represents not just an aesthetic but a body of lived experience, a way of seeing that was developed through years of attention and care, and that to reproduce it at scale and on demand is to hollow out the thing that made it meaningful. Miyazaki's comment about not knowing "what pain is" was not a legal argument. It was a statement about what he believes art is for.
That belief deserves respect from fellow creators, even those who don't fully share it. And it deserves something more than dismissal, which is often what it gets in the AI creativity discourse, where the legal argument, "style isn't copyrightable," is sometimes used as a conversation-stopper rather than a starting point.
At the same time, the counter-position deserves more than caricature. The people generating Ghibli-style images in 2025 were not, in most cases, trying to defraud anyone or claim a lineage they hadn't earned. They were doing what humans have always done with things they love: finding a way to participate in them. The impulse behind "Ghiblification" is not categorically different from the impulse behind fan art, which has existed for as long as there have been beloved artistic traditions, and which most artists regard with warmth rather than hostility. The question is whether scale changes the ethics, and the honest answer is that it might.
There is a version of this conversation that is purely reactive, focused on what AI can and can't do, what is and isn't legal, who wins and who loses. That conversation is happening everywhere, and it has its place. But the more interesting question for creators is a personal one: what does your own relationship to other artists' styles look like, and what do you want it to look like?
Using AI to work in a particular aesthetic direction isn't inherently different from any other form of influence, but the ease of it does remove some of the natural friction that has historically served as a check on how far you go. When developing a style takes years of study and practice, you tend to end up somewhere that is genuinely yours, because the process of absorption is also a process of transformation. When it comes to a prompt and reference images, you can stay closer to the source longer than any human ever could through effort alone, and that proximity is worth being intentional about.
The most interesting AI creators right now are not the ones asking "can I do this?" They're asking "what am I actually trying to make, and what is the relationship between my work and the work that shaped it?" These are age old questions, remade for a fundamentally new era. AI didn't invent them, it just made asking them more urgent, and more visible, than they have ever been before.
At OpenArt, we think about this a lot. We believe in creativity as a fundamentally human act, and we believe that the tools we build should serve that act rather than shortcut it. Part of what that means in practice is giving creators the ability to work with confidence about the IP status of what they make, which is why we built our Commercial Safety Check in partnership with Copysight. The legal questions around AI and artistic style are genuinely unsettled, and we aren't going to pretend otherwise, but what we can do is make it easier for creators to navigate them thoughtfully.
The conversation about inspiration, imitation, and where the line between them sits is one worth having openly, but without easy answers for now. It is, in the end, a conversation about what it means to make something, and that's a conversation that belongs to everyone who does.