A new version of the AI program ChatGPT just launched with a feature that allows users to create images in the style of Studio Ghibli. Given the studio and founder Hayao Miyazaki’s unique animation style and widespread fanbase, this feature immediately went viral. Users delighted in being able to turn their images into the fantastical styles of “Spirited Away,” “My Neighbor Totoro,” and “Kiki’s Delivery Service.”
The Trouble With AI
But like most things involving AI, this fun trend is steeped in ethical and copyright issues. These concerns are mainly about how the model was trained and whether it only used licensed and public-domain art. ChatGPT creator OpenAI is currently fighting several copyright lawsuits over this exact issue.
The problem is AI models need a lot of data, in this case, art, to learn from. Some artists or their estates are willing to license their work for this purpose. However, many artists refuse (if they are even asked) due to concerns over this technology trying to replace human artists. Cases like “Blade Runner 2049,” where the creator was asked, refused, and the company did it anyway, also exist. Because of this, many of these models are trained off stolen work.
OpenAI announced on March 25th, it would be taking a “conservative approach” in how it mimics artists. “We added a refusal which triggers when a user attempts to generate an image in the style of a living artist,” the technical paper states. Adding later on it “permits broader studio styles — which people have used to generate and share some truly delightful and inspired original fan creations.”
Leaving mimicking Studio Ghibli well within the bounds of this new protocol. Even OpenAI CEO Sam Altman got in on the viral trend and changed his profile picture on X (formally Twitter) to a Ghibli-style portrait.
The White House also used the feature to alter an image of Virginia Basora-Gonzalez crying while detained. A move that even supporters of the current president’s immigration reform found tasteless.
Was Ghibli Involved?
The company and Ghibli haven’t publicly commented on how the model was trained. Given past comments by Miyazaki, it’s highly suspect the images the AI learned from were licensed.
In 2016, Miyazaki, who is known for his whimsical hand-drawn style, was shown an AI demo and was “utterly disgusted.” The presenter explained to him that AI could “present us grotesque movements that we humans can’t imagine.”
“Every morning, not in recent days, I see my friend who has a disability,” Miyazaki explained. “It’s so hard for him just to do a high five; his arm with stiff muscle can’t reach out to my hand. Now, thinking of him, I can’t watch this stuff and find it interesting. Whoever creates this stuff has no idea what pain is.”
Adding he would “never wish to incorporate this technology into my work at all.” “I strongly feel that this is an insult to life itself.”
Unless Miyazaki has had a massive change of heart in private these comments are pretty clear. While he isn’t the only employee at Studio Ghibli, we doubt something this big would have been done without his knowledge.
If you want an example of an ethical use of AI, check out the “Star Trek” short “765874: Unification.” All likenesses and creative work that fed the AI behind that project were used with permission from the artist or their estate, and the results were phenomenal. AI, like most tools, is morally neutral and CAN be trained/used ethically, it just often isn’t.
