Since OpenAI unveiled its new "Chat GPT-4o Image Creation" function on the 25th, image generation using it has exploded. In particular, there are concerns that this model may have infringed on Ghibli's copyright as the image converted into the style of the famous Japanese animation production company "Studio Ghibli".
"It's fun to see people create images with ChatGPT, but (due to excessive use) GPUs are melting down," OpenAI CEO Sam Altman wrote on his X account on the 27th (local time). He said he would put a limit on the usage of the image generation function until it was made efficient. In fact, the rate of image generation began to slow down from the 27th.
Creating Ghibli-style images using ChatGPT is popular on major social media such as Facebook and Thread as well as X. Beyond using photos of yourself or your family, you are asked to redraw celebrities or historically famous photos in Ghibli style.
However, there are also growing concerns about whether such style learning has any copyright issues. Josh Weigensberg, a partner lawyer at the law firm Fryer Cashman, told the Associated Press that the AI model of OpenAI could be a problem depending on whether it was trained with the work of studio director Ghibrina Miyazaki, explaining that the question is "whether you have a license or approval to train."
Weigensberg added that if this type of work is being used without consent and compensation, it could be problematic. He said that "style" is not protected by copyright, but stressed that it cannot be completely free.
"This is a clear example of companies like OpenAI not caring about artists' works and artists' livelihoods," said artist Carla Ortiz, who is filing a copyright lawsuit against other AI-generated image companies. "This is to advertise (OpenAI's) products using Ghibli's branding, name, achievements, and reputation."
[Silicon Valley = Lee Deok-ju Correspondent]
