Ghibli Goes Viral & Not Everyone’s Happy

On March 25, ChatGPT did an image generation update. On the same day or the next, a tweet appeared featuring popular memes in the aesthetic of Ghibli Miyazaki. It was pretty clear the intention of the creations to resemble Ghibli. I mean, though, it’s not the same and I’m not an art or illustration expert. As soon as I noticed the trend, I was like, ok, is this a new filter? I then saw that those who use ChatGPT should be able to generate images on a whole different level, especially those who are subscribers.

I am a horror writer and Monday through Friday during office hours I’m a digital marketer, but my background is in Graphic Design and though I was never one of the better illustrators in my class, I developed a love and respect for the art. I share that as context when writing about my perspective on how this whole Ghibli takeover the internet came from.

Sophie, the main character from Studio Ghibli’s “Howl’s Moving Castle,” stands in front of a mirror adjusting a wide-brimmed hat with a red ribbon and flower. She wears a green dress with a long braid tied in a pink bow, surrounded by vibrant hats and rich decor in her hat shop.

Sophie in the Hat Shop – “Howl’s Moving Castle” (Studio Ghibli)

ChatGPT’s image update

ChatGPT’s update (GPT-4o) introduced the ability to create or transform images directly within the platform with simple text prompts. The improved output was the key factor in how high-quality ‘Ghibliesque’ images began trending, closest I’ve seen at least generated through AI.

Apparently OpenAI also relaxed some content generation restrictions and I can attest to that. Just when I could finally upload images, around the end of 2023 or early 2024, I remember testing ‘Ghibli like’ requests. Every single time ChatGPT turned me down. I tried other styles and every single time ChatGPT would say it couldn’t do it because of copyright infringement. Now, I’ve been thinking if OpenAI overlords saw so many requests from other folks besides myself, they thought it was necessary to relax such restrictions. That would not be surprising and at least for now we know ChatGPT doesn’t see our keystrokes. For now.

As soon as I saw that tweet on X with the popular memes with Ghibli’s style vibe, I knew two things. I knew I wanted to try it with family photos and I also knew that if it were to go viral; it was going to cause a storm, which it did. I didn’t expect for it to go viral.

The viral tweet

Ghibli style AI filters have been around for a bit, through several apps, including TikTok. I’d say trends and virality move faster so many times it’s challenging to see the moment it crosses platforms and worldwide.

I wanted to share the exact moment this turned viral, but it’s difficult to pinpoint. The most notable early example I found was this tweet from Grant Slatton, which as of today has over 45K likes and 51 million views.

Within the same day, during the evening, I’m pretty sure I was already sending my family tons of “Ghiblified” images. At home, we’d already been exchanging digitally modified pictures (using apps like B612, Snapchat, and Instagram). This Ghibli trend felt different. It was so contagious, like a virus. We couldn’t wait to see precious moments of our lives in this aesthetic. It didn’t feel wrong for a second, even though I admit it was a very self-serving thing. Maybe it’s a JC thing, but whenever I’ve seen almost every single Ghibli world, I’ve dreamed of being part of it and the convenience that ChatGPT presented in the moment was hard to resist.

Accessibility & Virality

So, yes, it is self-serving if I may dare say and if you disagree with me, it’s ok. We made it about us, which is what we humans usually do. It’s why I’m sure it is one reason it went viral. It became one of those challenges almost anyone and everyone could take part regardless of age, talent, and socioeconomic level. Some even called it ‘democratic’. As long as you had internet, everyone could join the trend, still can though it has died down.

Side-by-side comparison of an AI-generated Ghibli-style illustration of a walrus’s birthday and the real-life photo it was based on. Both show an 8-year-old walrus named Nikolai wearing a pink tiara, sitting in front of a birthday cake, with staff cheering through a window at an aquarium in Dalian, China.

I posted a single public image: a ‘Ghiblified’ screenshot from a video of a walrus’s birthday party at a Chinese zoo. I saw it and wondered what this Nikolai would look like, took a screen shot and ran it through ChatGPT. What compelled me to ‘Ghiblify’ it? Curiosity.

On social media, there are trends people feel foolish to participate in, but the ones that are easier to replicate make sense would get more ‘virality’. Accessibility and convenience together are a powerful duo.

And as of today, there’s another trend picking up, the ‘action doll box’. That one is also interesting, but I won’t be writing about it, at least for now.

Google Trends Data

Through Google trends, I wanted to have a more real visual idea of how this AI-generated style went and, as you can notice (graphically generated with stats from Google Trends on ChatGPT).

Three charts showing the spike in global interest for “Ghibli style” in late March 2025. Includes a timeline of the trend's peak, top 10 related Google search queries, and countries most interested — led by India, Sri Lanka, and Tunisia.

I wish I could pinpoint where the most controversial takes are on the Ghibliesque trend, but at least the top three countries interested in the trend (at least through the specific term and related queries) were India, Sri Lanka, and Tunisia.

I couldn’t find Japan in the top 40 countries showing interest in this search, and their peak at least through Google search on this topic shows as April 7 and not the March 30th that peaked worldwide. Interesting.

Line graph comparing US search interest from March to April 2025 for “Ghibli style,” “AI art,” “Taylor Swift,” and “Naruto.” Ghibli style trends briefly spiked but stayed lower than Taylor Swift and Naruto throughout.

Miyazaki’s View On AI

There’s this quote by Miyazaki Hayao on AI that keeps making its rounds: the translation reads “an insult to life itself”. The clip I keep seeing on social media is rather grim and I’m not sure it’s the best example of how the average person is using it for visuals.

I believe the clip is from the 2016 NHK documentary series by Kaku Arukawa. The video starts at his house, I think. Eventually they’re maybe at Ghibli’s offices, I’m not sure.

Kawakami Nobuo, Chairman of Dwango/Ghibli Production Intern, walks in and Miyazaki asks what this is about. Kawakami says that they’ve been quietly doing research on CGI, ‘deep learning’. He mentions he thinks computers will soon be able to paint like humans five to ten years from that moment. We’re in 2025, so I don’t think he was wrong.

Miyazaki replied, ‘If they do that, we won’t need humans’. as he smokes a cigarette and laughs with Kawakami. Though I’m not sure if it’s a nervous laugh or sarcasm. I can’t tell; I don’t speak Japanese and he covers his mouth with his hand so it’s difficult to get a read on this statement. He moves to a conference room.

Kawakami tells Miyazaki that they have started using AI for image processing and he would like to show him what they do.

This is when I get confused. I have formulated a conspiracy theory in my brain. Why? This is literally the first image they show in the clip on the TV screen that they’re demonstrating to Miyazaki.

Computer screen showing four angles of a grotesque, AI-generated humanoid crawling on a checkered floor. Subtitle reads: “We taught the program to speed up the image.”

Close-up of the AI-generated humanoid creature crawling on its head across a checkered floor. Subtitle reads: “It uses the head to move.”

Detailed view of the AI-generated figure’s back and limbs contorting unnaturally. Subtitle reads: “No sense of pain, no idea that the head is important.”

Kawakami keeps talking. ‘It uses the head as if it were just an ordinary leg. It’s creepy, and might be good in a zombie game. Using AI enables a kind of horror beyond human imagination.

I think this gentleman should read my book when it comes out. Maybe he will say ‘beyond human imagination, so perhaps JC isn’t human’. I know, I know a bad joke haha.

‘And that’s what we’re doing.’ The camera pans back to Miyazaki’s face and oh boy.

‘Let me just say. Every morning, I used to see a friend who’s disabled. He would walk up to me. One leg’s turned outwards, so it’s hard for him to walk. Even a high five is hard for him. His stiff hand and mine touch. I think of him, and can’t say I like this. Whoever made it gives no thought to pain. It’s very unpleasant. You can make horrible things if you want, but I want nothing to do with me. It’s an awful insult to life’.

My question is, would Miyazaki have had the same visceral reaction would the animation the team would was showing something nice, cute, wholesome? I doubt it would have been positive because even before seeing the horror video, you could tell he was going to have a rough take on AI-generated visuals.

Kawakami and the other men next to him are barely blinking as they hear Miyazaki say these things in a calm tone, rubbing his arm/elbow with his hand. It gave me goosebumps when I saw that scene.

I won’t transcribe the rest of the video, but it ends with them explaining this is all experimental, not something they’re showing publicly. The other senior next to Miyazaki asks what the goal of all of this is. Another man responds that it is for a machine to draw pictures like people do.

There’s a moment of silence. I can’t help but stare at the background behind these men. A white board with the written words ‘deep learning’ and then there is a cut, camera goes back to Miyazaki while he draws, he says ‘I fear the world’s end is near. Humans have lost confidence.’

I’ve seen the shorter version of this clip several times but now that I’m pausing, rewatching for screenshots, to transcribe some quotes, I got teary. Is AI really a blessing?

Final Thoughts

To ‘democratise’ access to knowledge and art. Is that really the goal or are we just craving to increase our output? Fast food art, fast food literature, fast food life. The smile I had first when I was deep into the ‘Ghiblification’ of my family and wedding photos has faded.

Am I being dramatic? I will not be hypocritical and call others names for using AI applications to produce things, certainly at work, that’s a given. It’s expected in marketing, after all. I wrote an article on Teabot.AI in ‘collaboration’ with ChatGPT. It was a fun experience to share with a friend who has truly embraced this AI era.

Yet, I’m so nervous of putting my horror manuscript through ProWritingAid even though I know all I want is editing suggestions for grammar, duplicates, to clean some noise, just as I’ve done with this article. I keep repeating myself that eventually, I’ll hire a human editor so the experience feels less lonely, and hopefully, their expertise and collaboration will make me a better writer when they help with my horror manuscript.

I’m not sure what to expect of the future. This innocent trend has stuck with me since it first came out. What does this all mean? I hope to see what the brain of two artists who I respect their art and their voices on this matter. Fingers-crossed next I get to share a brief article with their answers.

How do you feel about this trend? About Miyazaki’s words? Is AI art real art? Are you a Luddite yet?

Thanks for reading this much.

——

Interesting reading:

Hayao Miyazaki’s AI Nightmare

AI may be coming for anime, but Hayao Miyazaki is irreplaceable, son says

Sam Altman defends AI art after Studio Ghibli backlash, calling it a ‘net win’ for society



2 Comments

  1. Jim McCabe wrote:

    This is a fantastic essay – thank you, JC. Fascinating for many reasons, but I have to comment. You are right – this is a stampede to nostalgia. Less like the thrill of the hula hoop fad – more like the frantic rush to the first water truck after a natural disaster. That desperate grasp for nostalgia is a manifestation of forfeited human credibility. Though chronically entertained, we are, most of us now, aware of the datafication of our human experience into trackable digits. We feel like a number, yet we revel in self-adoration as if a star in our own movie. And while genuine human ability is now recycled as nostalgia fetish, we are all endowed with amazing abilities. We ARE special – and we know it in our bones that we are manipulated, abbreviated versions of what we could be, what we aspire to be. That’s why AI is so horrifying to Miyazaki – as it should be to all of us. Just like a parrot can’t really hold a conversation, AI can mimic and assemble composites, but it can’t really know what it’s like to bite into a juicy peach on a hot summer day; to be slapped in the face by a friend; to say goodbye to someone you love. It’s the lived experience that authenticates creativity – our inherent, species-specific gift. In this faithless moment in humanity, we reach for nostalgia to touch these feelings – even if it’s AI-generated fluff. Who among us wants to say that feelings are not relevant? Industry. Business. That’s who. AI is a productivity tool that replaces humans like no other tool we have created. And all the amazing and fun things you can do with AI perfectly serve these AI Evangelists, who make a familiar promise: I will set you free – just relinquish your soul to thee. Slavery is bad, no matter how entertaining, benevolent or unfeeling the master. Just words from a guy from another century.

  2. JC wrote:

    We feel like a number, yet we revel in self-adoration as if a star in our own movie.
    You speak my language, Jim.

    Your point about “AI setting us free” reminded me of the hypothesis about how cooking with fire helped humanity evolve. By cooking, we had more time and energy to do other, more important things. Part of that hypothesis suggests that having more time allowed early humans to think, to philosophise/wonder.

    I don’t see AI yet in the same way, because in some cases it feels like it might be taking the “thinking” away. Let’s hope I’m wrong. Either way, we’ll find out in a few years.

    Thanks for reading my article and for sharing your thoughts, Jim!

Leave a Reply