New NHK Video Report Shows Generative AI Use in Anime/Manga Industries – Interest
The report expands on an August Nikkei newspaper article
©NHK (Japan Broadcasting Corporation) All rights reserved.
The NHK World-Japan YouTube channel released a report on the use of generative AI in the anime and manga industries on December 9. In the video’s description, NHK World-Japan said, “Manga and anime are a big part of Japanese culture. But they require a huge amount of work to produce. Now, AI is taking on some of that work, so humans can focus on the creative side.”
https://www.youtube.com/watch?v=CTwGwF7cw8A
The report first focuses on 70-year-old manga creator Yoshimi Kurata and the production of his manga. Kurata notes his drawing speed is slowing down and with the use of the technology, he can continue producing manga for another 10 years as opposed to five.
The report then focuses on the Nagoya-based animation studio K&K Design and their use of generative AI in production. The report shows how the studio creates in-between animation frames with the technology, but has human staff members check and fix all material produced with generative AI. A staff member interviewed for the NHK piece notes the in-between animation frames normally take one week to 19 days, but generative AI shortens that to four to five hours.
Hiroshi Kawakami, a director at K&K Designsaid in the report:
We always make sure a human checks, retouches, or adds to the work. As creators, we don’t want to rely too heavily on generative AI. However, we believe AI can save significant time, so we can use that time for more creative things.
The Nihon Keizai Shimbun (Nikkei) paper published a similar report in August. The Nikkei article also focuses on K&K Design as well as Tokyo-based digital manga production company en-dolphin (pronounced endorphin) and localization company Orange Inc. (no relation to animation Studio Orange).
As generative AI is still a developing technology, the definitions and ethical use is still undefined by many. Anime News Network reached out to Dodzi Dzakuma, a Senior Azure Cloud Engineer and Software Developer, regarding a definition of generative AI. Dzakuma said:
Generative AI isn’t actually AI per se. It’s more of a set of branching “if then” statements that allow a model trained on set of data to accurately use human language to produce the most likely tangible outcome base on a decision tree. We collectively have called it AI, but they’re decision trees that generate output.
Dzakuma did not comment on the ethical use of the technology.
Sources: Email correspondence, NHK World-Japan’s YouTubechannel