Around a day after the release of image generation in ChatGPT, OpenAI CEO Sam Altman announced that the feature would not be rolled out to free users as promised, citing its popularity.

“[I]mages in chatgpt are wayyyy more popular than we expected (and we had pretty high expectations). rollout to our free tier is unfortunately going to be delayed for awhile,” posted Altman on X on March 27.

Users could previously generate crude AI images with OpenAI’s DALL-E text-to-image generation models, but the results were often far from impressive and contained multiple jarring errors or distortions.

However, image generation in ChatGPT attracted new users soon after release when Altman and other early users shared pictures of themselves converted into an anime style.

Other users on X then began turning their own photos, portraits, or family pictures into cute Studio Ghibli-style images, which the model was able to mimic to a remarkable degree. Studio Ghibli co-founder Hayao Miyazaki’s lush and expressively animated films such as ‘Spirited Away’ and ‘Howl’s Moving Castle’ have fans world over; ChatGPT now allows those incapable of creating original art to generate Miyazaki’s style natively with just text prompts and refinement.

While Miyazaki had expressed disgust and anger over AI-generated animation and art several years ago, this did not stop fans and anime watchers from turning memes, Bollywood images, and even crime scene photos into pictures that looked as if they had come from the iconic director’s family-friendly films.

Even though other AI models can sometimes generate Studio Ghibli-style images on the basis of existing photos, these are often poor imitations that do not look much like the original content. Other AI models refuse to create such images due to their content filters. This is one reason why image generation in ChatGPT has captured the public imagination.

Altman himself updated his X profile picture to show an anime version of himself that made him look like a Studio Ghibli character.

“[B]elieve it or not we put a lot of thought into the initial examples we show when we introduce new technology,” posted Altman on X, in response to a user who approved of the use of Studio Ghibli aesthetics to introduce image generation in ChatGPT.