Interesting thing about Japanese copyright law is that if they deem that the generated AI was used for "enjoyment" you are still in violation of the law. Because Japan gave AI companies to commercialize AI-generated works, some people think that they can do whatever they want, but that's not really the case - in some ways, it's actually more strict than laws in the US.

So for OpenAI, they have several problems on hand right now:

Interesting thing about Japanese copyright law is that if they deem that the generated AI was used for "enjoyment" you are still in violation of the law. Because Japan gave AI companies to commercialize AI-generated works, some people think that they can do whatever they want, but that's not really the case - in some ways, it's actually more strict than laws in the US.

So for OpenAI, they have several problems on hand right now:

https://www.privacyworld.blog/2024/03/japans-new-draft-guidelines-on-ai-and-copyright-is-it-really-ok-to-train-ai-using-pirated-materials/

- AI art is a business model that undermines itself. The easier it is to generate something, the more you will get of it, oversaturating the market, eventually leading to its own collapse. The need for innovative and unique ideas from humans will always be needed but that's not really what these models are designed for. We don't even need to talk about quality because this is just a matter of supply and demand.

- Due to OpenAI's loss to China and the generally disrespectful tone that AI companies take towards the creative class, there's not much value in "owning" AI art because its prestige value is mostly negative at the moment.

Personally I don't see any of these scenarios ending well for OpenAI - they're really in over their heads at this point. 🌊

- If they audit their data/comms and find that they've been *willingly* training on copyrighted material, they are basically done. This is what makes Suchir Balaji death even more strange, given that AI companies are still largely sitting on a house of cards, legally speaking.

- We've yet to hear from Miyazaki himself about this particular issue, but it's safe to assume that he probably has some skepticism about the tech industry's "dehumanization", which he has been very vocal about in the past. If he argues that AI-usage of his style is "harmful" to Ghibli's branding, they will be forced to remove it from their model, at least in Japan. (In US, we have Watterson's Calvin and Hobbes which is strictly enforced - which is why you never see that style pop up too much, even in tech products.)