Japan asks OpenAI to keep Sora 2's hands off anime IP | Hacker News Reader

Japan asks OpenAI to keep Sora 2's hands off anime IP

theregister.com

63 pointsmaxloh19 comments

> "When I tried entering a prompt into Sora 2, it generated a succession of images of popular anime characters with such high quality that it was indistinguishable from the real thing," Shiozaki said in a blog post written days after Altman's Sora 2 note. "However, for some reason, characters whose rights are owned by major American companies, such as Mickey Mouse or Superman, did not appear."

So they know what they’re doing is wrong and they’ll just abuse the people who can’t sue them as easily?

Yeah that tracks.

It is indeed a vile practice, and it works every time. More money to them and almost complete impunity. Talk about a business model.

Yeah, that's Sam's MO -- overpower and abuse the little people.

Like his sister…

To be fair it probably is possible to fully exhaust the possibility space for reasonably human looking anime style characters with a few hundred gigawatts of compute.

Just by generating every possible combination of features within human drawing limits.

So copyright as a concept might not necessarily be coherent in the future, at least for anime characters.

This has already been tested with people generating mass amounts of text. The conclusion was that fully automated generation doesn’t count. There has to be some human curation and creativity applied.

Why wouldn’t it count as prior art for any future work?

When I see people with Ghibli style profile pictures it's an immediate signal they have no taste.

If copyright industry is going to win this battle then it means nothing is going to stop their madness. The only way forward is abandoning human created content altogether same way open source software abandoned proprietary code and tools. Funnily enough AI output filtering that strictly removes any products of copyright industry might help with that.

Do some artists want tools to enable their vision faster? Or is there more value in slow art that reaches the same endpoint?

Yes, there are people everywhere who want to take shortcuts in life and do as little work as possible. They'd agree with the implication that this is only a matter of speed, they'd think it's plausible that they'd reach the same endpoint with profoundly different methods. They'd make up whatever justification necessary.

Not everything needs to be a race, and certainly not art.

> Do some artists want tools to enable their vision faster?

Artists are expensive. The goal is to eliminate them ( just like with SWE).

"Gemini, generate a Picasso from those images. "

Why not do it in the reverse - Japan has shortage of animators, why not use Sora instead to help them and increase productivity?

It's the creators' choice if they want to go that way. They could've used pose interpolation for quite a while. Going as far as Sora for it is an overkill. That doesn't solve the IP stealing issue in any way though.

It will get stolen after animation is released anyway. So kind of moot point?

Not for creators taking a stand. Futile actions are sometimes important to people.

So you get security to guard you warehouse from theft. Which is good.

But why not do it in reverse and buy stolen goods!

Reply on news.ycombinator.com