Social media users are creating Studio Ghibli-style portraits using ChatGPT’s new native image-generation capabilities.
The big picture: The Ghibli fest is dividing admirers of the Japanese animation studio, with some awed by the visuals and others dismissing them as “AI slop.”
Driving the news: OpenAI on Tuesday rolled out the images feature in GPT-4o, initially making it available to users on the Plus, Pro, Team and Free subscription tiers for creating or adjusting images. It’s also available to use in Sora.
- “We trained our models on the joint distribution of online images and text, learning not just how images relate to language, but how they relate to each other,” OpenAI said in a blog post.
- OpenAI stated that it will “block requests for generated images that may violate our content policies, such as child sexual abuse materials and sexual deepfakes.”
Between the lines: The Ghibli-styled images are possible because OpenAI has loosened rules that used to limit the creation of images using distinctive styles, alone with some celebrity likenesses, company names and logos.
- ChatGPT still won’t “generate an image of Taylor Swift” but it will offer to match a celebrity’s “vibe” or “aesthetic.”
State of play: Hundreds of users across social media platforms are using the native image feature to convert old photos to Studio Ghibli-style images.
- One person on X said: “tremendous alpha right now in sending your wife photos of yall converted to studio ghibli anime.”
- A person on Threads said that he was in “love with ChatGPT’s new image editing feature” and that he “can turn all my family photos into Ghibli portraits.”
- Another on Bluesky recreated stills from the “Star Wars: The Last Jedi” movie.
Yes, but: Not everyone is pleased.
- One Threads poster said “shame on you” if you’re using AI to generate “Studio Ghibli looking slop.”
- Another person said: “Some of you all ain’t ever got a Studio Ghibli copyright strike and it shows. Also Ai Art is SLOP.”
Go deeper: AI’s promised nirvana is always a few years off
