OpenAI’s viral Studio Ghibli moment highlights AI copyright concerns
It was only a day since the new AI image generator of Chatgpt went live, and social media-feeds have already been flooded with AI-generated memes in the style of Studio Ghibli, the cult-favorite Japanese animation studio behind Blockbuster films such as “Me Nighbor totoro” and “.
In the last 24 hours we have seen AI generated images that Studio-Ghibli versions of Elon Musk“Mr. van de RingsAnd President Donald Trump. OpenAi CEO Sam Altman even seems to have made a picture of the studio-in-style in the style of his new profile photoProbably made with the native image generator of GPT-4O. Users seem to upload existing images and photos to Chatgpt and ask the chatbot to make it again in new styles.
The newest update from OpenAi is on the heels of Google’s release of a similar AI image function in its Gemini Flash model, which also led a viral moment earlier in March when people used it to remove water brands from images.
The latest tools from OpenAi and Google make it easier than ever to re-create the styles of copyrighted works-just by typing a text prompt. Together, these new AI image functions seem to be concerned about the core of different lawsuits against generative AI model developers. If these companies train on copyright protected works, do they violate copyright legislation?
According to Evan Brown, an intellectual property lawyer at the Neal & McDevitt law firm, products such as the Native image generator of GPT-4O are working in a legal gray area today. Style is not explicitly protected by copyright, according to Brown, which means that OpenAi does not seem to break the law, simply by generating images that look like studio -Ghibli films.
However, Brown says that it is likely that OpenAi has reached this parable by training his model on millions of frames from the Ghibli films. Even if that was the case, various courts still decide whether the training of AI models on copyright protected works falls under the protection of reasonable use.
“I think this raises the same question that we have been asking for a few years,” Brown said in an interview. “What are the implications of copyright infringement on going out, crawling on the internet and copying to these databases?”
The New York Times and various publishers are in active lawsuits against OpenAi and claim that the company has trained its AI models on copyrighted works without the correct attribution or payment. Similar claims have been established in lawsuits against other leading AI companies, including Meta and AI-Startup Midjourney.
In a statement to WAN, a openi spokesperson says, while Chatgpt refuses to replicate ‘the style of individual living artists’, allows ‘wider studio styles’ to replicate. It is of course worth noting that there are living artists who are credited for pioneering the unique styles of their studio, such as co-founder of Studio Ghibli, Hayao Miyazaki.
Apparently users have also been able to use the function of the Native Image generation of GPT-4O to re-create styles of other studios and artists. Someone else made a Marc Andreessen portrait in the style of Dr. SeussAnd a couple has created them again Wedding photos in the Pixar style.
We have tested several popular AI image generators – including those available in Google’s Gemini, Xai’s Grok and Playground.ai – to see their ability to match the style of Studio Ghibli. We found the new image generator of OpenAi made the most accurate replica of the style of the animation studio.
For the time being, the new image functions of OpenAi and Google are presenting a leap forward in what AI models can generate, which seems to lead an increase in use. OpenAi has delayed the rollout of his new image tool on Wednesday, with reference to a high question. That is perhaps the most important thing for these companies today, but we will have to wait until the courts weigh their legality.
