Fans are using “Images for ChatGPT,” recently released by OpenAI, to remake personal photos and historical scenes in a look inspired by Hayao Miyazaki’s soft and whimsical approach. However, the trend is turning into strange territory, with some users applying it to dark or sensitive moments in history and to well-known public figures.
When OpenAI added image-generation features to ChatGPT, it showcased the system’s power to handle shadows, depth, and even readable text. Soon after, countless people were experimenting by feeding the tool prompts referencing Studio Ghibli’s iconic style.
Everyone from couples to pet owners began “Ghibli-fying” loved ones, resulting in cartoonish reimaginings that mirror the cozy, hand-drawn look fans associate with My Neighbor Totoro or Spirited Away.
Some folks compared it to the human-drawn art commissions sold on sites like Etsy, where artists offer “anime-fied” caricatures.
ChatGPT Ghibli creations have taken a weird turn
Once playful experiments started to involve scenes like the 9/11 Twin Towers attack, the assassination of President John F. Kennedy, and even Nvidia CEO Jensen Huang in bizarre scenarios. One widely shared set of images shows a Ghibli-style re-creation of President Donald Trump posed with Jeffrey Epstein.
While it remains unclear whether the system is pulling from transformed uploads or user prompts, Altman himself seems amused by the phenomenon. He has changed his profile photo on X to a Ghibli version of himself and is urging followers to submit fresh renditions.
Images for ChatGPT guidelines contain loopholes.
The GPT-4o system card states that the model “can generate images that resemble the aesthetics of some artists’ work when their name is used in the prompt.” The same card claims that a refusal should trigger if a prompt tries to replicate the style of any living artist, yet Miyazaki is living, and his look remains fully accessible for these AI outputs.
OpenAI spokesperson Taya Christianson said that the company is not blocking the model’s capability to depict adult public figures outright, though there are safeguards for photorealistic uploads of individuals. “Depicting violence in artistic, creative or fictional contexts is generally allowed to enable creative and artistic endeavors,” Christianson explained.
In an email, she added, “Our goal is to give users as much creative freedom as possible. We continue to prevent generations in the style of individual living artists. We do permit broader studio styles, which people have used to generate and share some truly delightful and inspired original fan creations.”
Despite that, many people appear to be transforming images they do not own.
“Users should only upload materials to our services that they own or have the rights to use,” Christianson said in the same statement.
This warning has not deterred those who have shared edgy or disturbing creations. Ghibli-style Hitler in Paris and images of Korean-American business owners holding guns during the 1982 Los Angeles riots have also made rounds online.
There’s an entire website dedicated to AI creations of public personalities in cartoonish Ghibli form.
outside of midjourney, when people would think of “ai generated images” they would mostly think “slop” but that all changed with a chatgpt update and one guy’s post about sending ghiblified images to his wife pic.twitter.com/UEYk7hD25G
— sophie (@netcapgirl) March 26, 2025
ChatGPT has some evident blocks in place, including one that declines direct requests for photos of public figures or popular cartoons. Yet it’s easy enough to bypass the restrictions, for instance, by describing general traits or referencing a “Totoro-like character,” which can guide ChatGPT toward the same end result.
The guidelines themselves mention that users must not infringe on anyone’s rights and suggest that people can report copyright violations. OpenAI states that repeated infringements may lead to account termination.
Beyond the issues around these Ghibli transformations, OpenAI has faced a broad range of copyright challenges.
Actor Scarlett Johansson once threatened legal action because a voice tool mimicked her speech patterns. Lawsuits from authors, media outlets, and artists have also targeted OpenAI, claiming the company used their work to train the models without proper licensing.
While these legal battles continue, OpenAI has kept its training data sources fairly secret. It argues that AI-generated content is fundamentally new and not merely copied, though opponents say the line is blurry.
Studio Ghibli and Miyazaki aren’t the only creative names in the AI takeover
Users have pushed ChatGPT to produce images reminiscent of Rick & Morty, Wallace and Gromit, The Simpsons, and South Park.
There is no clear indication whether the teams behind these shows plan to object to the freewheeling AI references or if they might dispute the source data used in training.
For now, people who prompt ChatGPT to mix, match, or recreate their favorite show’s visuals are typically left to do so without formal complaints.
Yet Miyazaki’s own stance on AI-driven imagery is well documented. Back in 2016, before mainstream generative AI took hold, he was shown a “deep learning” clip at a production studio.
— tomie (@tomieinlove) March 26, 2025
He reacted strongly, calling it “an insult to life itself” and saying, “I am utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it, but I would never wish to incorporate this technology into my work at all.”
Cryptopolitan Academy: Tired of market swings? Learn how DeFi can help you build steady passive income. Register Now
No comments yet