OpenAI’s recent image-generation update has swept social media, inundating platforms with beautiful Ghibli-style art. From idyllic watercolor scenery to emotive hand-drawn characters, the internet now looks like something out of Spirited Away and Howl’s Moving Castle.

How Did the Ghibli Trend Start?

Launched on Tuesday as part of OpenAI’s GPT-4o upgrade, the new image-generation tool has amazed users with its ability to replicate the beloved Studio Ghibli aesthetic. While the AI was designed to produce highly detailed and stylized visuals, its unexpected mastery of Hayao Miyazaki’s iconic animation style has turned it into an online sensation.

Related News

Shortly after the update rolled out, users began experimenting with different prompts to test the AI’s artistic capabilities. They quickly realized that the model excelled at creating Ghibli-style art. As a result, social media platforms like X (formerly Twitter), Instagram, and Reddit were soon inundated with AI-generated Ghibli-inspired images.

Users posted renditions of their selfies, pets, fictional characters, and even mundane objects such as coffee cups and bicycles, all converted into fantastical Ghibli-style masterpieces. The trend picked up steam under hashtags such as #GhibliStyle and #AIGhibli, with some users even reimagining celebrities, historical figures, and internet memes in the Ghibli style.

OpenAI CEO Joins the Trend

Even OpenAI CEO Sam Altman got into the viral act. On Wednesday, he jokingly commented on X, “Be me. Grind for a decade trying to help make superintelligence to cure cancer or whatever, mostly no one cares for the first 7.5 years, and then for 2.5 years everyone hates you for everything. Wake up one day to hundreds of messages: ‘look I made you into a twink Ghibli style haha.'”

Altman himself even changed his profile picture to a Ghibli-style AI-made version of himself, exacerbating the trend.

ChatGPT 4o’s Image Generation Revolution

OpenAI’s latest ChatGPT update, powered by GPT-4o, now enables users to generate images directly within the chatbot. This feature, called Images in ChatGPT, is available to all users, though free-tier users face limitations due to high demand.

But the deployment wasn’t perfect. OpenAI briefly limited free users’ access and had to deal with a glitch where the AI could create images of “sexy men” but not “sexy women.”

The new model, despite small problems, is a significant advance in AI-generated visuals. In contrast to the earlier models like DALL-E, which created images in their entirety at once, GPT-4o builds images in stages, with better accuracy, text rendering, and object relations.

With Ghibli-style art created by AI taking over social media, this newest development in picture-making is one of the most culturally significant updates to date from OpenAI.