ChatGPT's image generator sets off a Ghibli wave

Within a day of OpenAI switching on native image generation in GPT-4o, ChatGPT users began flooding social media with photos and film stills converted into the visual style of Studio Ghibli, the Japanese animation house behind Spirited Away and My Neighbor Totoro. Sam Altman changed his own profile picture to a Ghibli-style image and, as the trend peaked, posted that ChatGPT had added a million new users in a single hour — against the roughly five days it took the original ChatGPT to reach the same figure in late 2022. Demand was heavy enough that OpenAI imposed temporary rate limits and delayed the feature’s rollout to free-tier accounts, with Altman writing that “our GPUs are melting.”

The trend reopened an argument that had run since the first image-generation lawsuits: whether recreating a studio’s recognisable visual style, as opposed to copying a specific image, infringes copyright. Commentators noted that style itself sits in a legal grey area — intellectual-property lawyer Evan Brown was quoted saying style is not explicitly protected — even as The New York Times and other publishers pursued separate, active lawsuits against OpenAI over the use of copyrighted material in training data. OpenAI said its policy was to decline requests to mimic the style of individual living artists while permitting broader studio styles, a distinction critics said was hard to sustain given how closely the Ghibli look is identified with its co-founder Hayao Miyazaki, who had said in a widely recirculated 2016 documentary clip that he found AI-generated animation “an insult to life itself.”

The episode became a reference point in the copyright debate less for any legal resolution — none followed directly from it — than for demonstrating how quickly a capability upgrade could turn into mass replication of a specific, recognisable artistic style at consumer scale.