ChatGPT has sparked a viral trend with its new ability to generate images resembling distinctive styles, with Studio Ghibli‘s aesthetic capturing widespread attention. After OpenAI released the image generation feature to ChatGPT, users worldwide began transforming personal photos and creating artwork reminiscent of beloved classics like “My Neighbor Totoro” and “Spirited Away” with ChatGPT and other AI chatbots as well.
Social media platforms have been flooded with these AI-generated Ghibli style images, with even OpenAI CEO Sam Altman participating by changing his profile picture to a Ghibli-inspired portrait. But how does ChatGPT mimic an art style that takes human artists years to master? The answer lies in GPT-4o’s approach to image generation.
This is how ChatGPT is creating Ghibli style images
The image generation feature of ChatGPT is powered by GPT-4o’s advanced multimodal capabilities, enabling the chatbot to process text prompts and generate corresponding images with precise detail and artistic flair.
Unlike previous AI image generators like DALL-E 3, GPT-4o employs an autoregressive approach to image creation. This method constructs images pixel by pixel in sequence rather than refining them from noise, resulting in more detailed and coherent outputs. The model’s ability to understand context from ongoing conversations also allows users to iteratively refine their images through natural dialogue.
This update represents a significant evolution not just for ChatGPT but for artificial intelligence systems as a whole. GPT-4o’s architecture integrates text, image, video, and audio processing into a single system, enabling seamless transitions between different media formats and more intuitive user interactions. Beyond Ghibli style images, users can create art in various styles and even generate photorealistic pictures.
To some extent other chatbots like Gemini and Grok can also create images reminiscent of Ghibli style artwork, however the resemblance of ChatGPT remains the most uncanny as of now.
Is ChatGPT allowed to create Ghibli style images
The popularity of Ghibli style image generation has reignited debates about AI art and intellectual property rights.
OpenAI has not disclosed whether Miyazaki’s works were specifically included in GPT-4o’s training data, raising questions about the ethical implications of the model’s ability to so closely mimic Studio Ghibli’s distinctive style. The company claims that GPT-4o includes safeguards against mimicking living artists’ styles, though the replication of Studio Ghibli’s aesthetic continues to raise complex ethical questions.
Studio Ghibli co-founder Hayao Miyazaki has previously expressed strong opposition to AI in creative processes, describing it as “an insult to life itself.” Critics argue that training AI on existing artworks without explicit consent potentially undermines artists’ livelihoods and creative control.
Legal experts note that while artistic styles themselves aren’t typically protected by copyright, the reproduction of specific elements from copyrighted works could potentially cross into infringement territory.
OpenAI emphasizes that the ChatGPT’s image generation model’s purpose is to augment human creativity rather than replace it.
