Toshio Suzuki dismisses AI 'Ghibli-style' art

Toshio Suzuki’s comments on AI-generated “Ghibli-style” art did not originate in a new statement on social media on Tuesday. They came from an interview with Japan’s Asahi Shimbun published on September 6, and were then recirculated in social posts and follow-up news reports over the next two days. The Studio Ghibli co-founder and chairman said the flood of AI images modeled on the studio’s look had “neither meaning nor fun,” according to Korean and English-language reports that cited the Asahi interview. He also said AI could imitate surface features but not what he described as the studio’s actual sensibility. The repost that circulated on X on September 8 appears to have amplified those earlier remarks rather than revealing a fresh on-the-record comment from Suzuki that day. (khan.co.kr) The underlying reporting traces back to the September 6 interview. ### Where did Suzuki’s remarks actually come from? Asahi Shimbun published the interview on September 6, according to multiple secondary reports that quoted it. In that interview, Suzuki was asked about the wave of AI-generated images that mimicked the visual style associated with Studio Ghibli. (khan.co.kr) Kyunghyang Shinmun reported that Suzuki said, “Some kinds of animation can be made easily with AI, but if you have a certain level of knowledge, you will see right through all of it.” The Herald Business separately quoted him as saying that “anyone with even a little knowledge can immediately tell it’s fake.” (khan.co.kr) ### What, exactly, did he say about the AI images? Suzuki said the images had “neither meaning nor fun,” according to Kyunghyang Shinmun’s English report. (khan.co.kr) The same report said he added, “AI cannot imitate Ghibli.” The Herald Business framed the same point in slightly different language, saying Suzuki believed AI might reproduce the outward appearance of a “Ghibli style” image but could not reproduce the studio’s “genuine expression and sensibility.” That distinction — surface resemblance versus underlying expression — is the core of his criticism in the available reporting. (khan.co.kr) ### Why did he say so little when the trend was at its peak? (khan.co.kr) Suzuki said he turned down interview requests when the AI-image boom was at its height. According to both Kyunghyang Shinmun and The Herald Business, he said he believed public attention would fade quickly. The Herald Business quoted him as saying, “I thought everyone would get bored of it quickly, so I declined, and that’s exactly what happened.” Kyunghyang Shinmun reported a similar line: “There were many interview requests, but I thought everyone would get bored quickly, so I turned them all down.” (biz.heraldcorp.com) ### Was this tied to the earlier OpenAI image boom? Kyunghyang Shinmun said the trend accelerated in March last year when OpenAI released an image-generation model that made “Ghibli-style” prompts widely visible online. (khan.co.kr) The report noted that users uploaded portraits and asked for them to be transformed into images resembling characters from films such as *Spirited Away* and *My Neighbor Totoro*. That report also said OpenAI Chief Executive Sam Altman posted at the time that GPUs were “melting” as usage surged. (khan.co.kr) Suzuki’s comments, as reported this week, were a retrospective response to that broader wave rather than a reaction to a single new product launch on September 8. ### How did Suzuki connect this to Hayao Miyazaki? The Herald Business said Suzuki contrasted AI’s data-driven production with Hayao Miyazaki’s working methods. (khan.co.kr) It reported that Suzuki described Miyazaki as creating in near-total isolation during production of *The Boy and the Heron*, without a smartphone, newspapers or television. Suzuki said that process produced imagery that computers struggle to reproduce, according to the report. (khan.co.kr) The same article said he pointed to a scene from *The Boy and the Heron* that experts had described as something computer graphics could not achieve. ### What is the clearest takeaway from the September 8 repost? The clearest verified point is that Suzuki’s remarks were real, but not newly issued on September 8. (biz.heraldcorp.com) They were reported from an Asahi interview published on September 6 and then spread more widely through reposts and follow-up coverage on September 7 and September 8. As of Tuesday, the publicly circulating version of the story was still anchored to that interview, with Suzuki’s quoted criticism centered on two claims: that AI-made “Ghibli-style” images are easy to spot as fake, and that they lack the meaning and appeal he associates with Studio Ghibli’s work. (biz.heraldcorp.com) (khan.co.kr)