As OpenAI launched the Ghibli-style AI image-generating feature in ChatGPT, the average weekly users crossed the 130 million mark for the first time in 2025, creating more than 700 million photos. 

According to SensorTower data, with the integration of image generation capabilities in the GPT-4o model, the AI platform experienced an all-time high in active users, in-app subscription revenue, and app downloads.  

Sam Altman, OpenAI CEO, posted on X,

“We added one million users in the last hour”

However, this massive hit led to glitches and low-scale outages due to traffic spikes. For that, the co-founder of OpenAI stated 

“We are getting things under control, but you should expect new releases from OpenAI to be delayed, stuff to break, and for service sometimes to be slow as we deal with capacity challenges”

Studio Ghibli

Studio Ghibli is a Japanese animation film studio founded in 1985 by Hayao Miyazaki. The company is renowned for its high-quality filmmaking with hand-drawn animation and rich storytelling. Some popular movies by the studio are Neighbor Totoro, Spirited Away, Kiki’s Delivery Service, Howl’s Moving Castle and Princess Mononoke. These movies are remembered by viewers for their soft colors, aesthetic landscapes, heart-touching stories, and visually stunning animation. 

Ghibli-Style Gone Wrong

Though OpenAI succeeded in attaining global attention through the launch of a Ghibli-style image-generating feature, it is not for the right reasons. Multiple compilations of incorrect Studio Ghibli-style images have become viral on Instagram, with users saying ChatGPT generated an extra hand, extra child, and even two husbands from one photo. One user on Instagram noted, 

“Shocked and funny, had to post it”

Ghibli Copyright Controversy 

As the Ghibli-style images went viral, controversy over the Ghibli studio’s copyright and creative ownership also sparked. The AI chatbots like ChatGPT can be simply explained as style engines. These systems don’t store information in any traditional sense. Instead, they encode text, facts, or image fragments as patterns or styles. It means everything, including objects, writing genres, and voices, is transformed into styles. 

When ChatGPT learned of Miyazaki’s work, it did not store actual Studio Ghibli frames. Instead, it encoded “Ghibli-ness” as a mathematical pattern: a style that can be applied to new images. It is also one of the reasons that the AI Chatbot is going wrong with producing extra hands or extra people. When a prompt is given to the chatbot to convert an image into a Ghibli-style image, it only applies specific Ghibli-style features like soft pastel colors and aesthetic landscapes. However, the artists allege that their artwork is used to train these chatbots and make copyright claims. 

Amid this controversy, Miyazaki’s comments resurfaced from 2016 when he said 

“If you want to create the creepy stuff, you can certainly do that, but I would never want to integrate this technology into my works at all. I very strongly feel that this is an insult to life itself.”

Reverse-engineering of Ghibli-style Images

Delete your Ghibli filter pics ASAP. Someone can reverse-engineer the viral Ghibli filter and discover your original photo. 

This trend went viral on social media, sparking another controversy about privacy breaches. However, from a coding perspective, it is not possible. The filter creates a new image inspired by the uploaded image. The original image is not ‘inside’ the anime image. Though AI platforms take the consent of users before uploading an image, it is unlikely that a publicly uploaded Ghibli-style image could be reconverted to the original image. 

Google Opened the Beta Gemini 2.5 for Free Users, Competing with OpenAI

After ChatGPT made a sensation in the AI world with its free launch of Ghibli-style images, Google also opened its beta Gemini 2.5 for free users that can generate AI images. Although Google refrained from commenting on whether the chatbot can create Ghibli-style images or on which external systems it depends on visuals. However, after practical experience, it becomes clear that Gemini 2.5 Pro depends on Google’s Imagen 3 for image creation. According to a reviewer, Aman Gupta 

‘‘I made several other attempts to get Gemini to generate ‘Ghiblified’ images, using Grok and ChatGPT for prompts, but Gemini 2.5 Pro did not provide the desired image, with results ranging from a straightforward error message of some kind to the chatbot being left at the loading stage.’’ 

Ghibli-Style Will Fade Away. What’s Next for OpenAI? 

Though Ghibli-style images created a sensation with ChatGPT getting overloaded, such trends fade away with time. OpenAI is taking measures to handle such large traffic, however, it needs to reconsider its strategy, considering that every new day, there cannot be a user overload. Therefore, the AI platform needs to work on enhancing the quality of the chatbot along with the number of users. For instance, the platform could focus on making the Ghibli-style images more accurate. By then, it will continue to compete with the emerging AI chatbot DeepSeek, attracting users and investors with open-source modeling and cheaper production.