You must log in or register to comment.
It hasn’t though? That short “film” in the thumbnail is very not great. Just to summarize a few issues clothing items randomly change colors, clothes randomly switch to different clothes, they attempted to copy the really nice crane shot in Frieren but used “AI” to do it which just smeared ugly green shit around the screen cause there were leaves, a path they are walking on randomly changes material multiple times, ect ect. All the usual AI video crap. They didn’t solve any of the issues it has and it there’s one “animator” credited.
Spoilers if you care (no I’m not going to actually mark them) but there’s a part where using AI would have been a really good directing decision, especially with the faults but because the “director” is just a tech bro who’s incapable of understanding art they completely missed an extremely good use case.
Actual animation studios don’t and won’t use this crap until it’s dramatically better. The closest thing they currently use is cacani (actually an interesting tool scroll through that) and it uses absolutely no “Ai” whatsoever
For all of the quality complaints about this anime, we have to remember that the technology is improving at a breakneck pace. What we are seeing there is the state of the technology from over a year ago. They used Stable Diffusion, which barely anyone even uses these days, because it’s been left in the dust. It was also an image generation model, which is what caused most of the issues that the anime had–the model was never designed for use on video in the first place. But now we DO have video models, which can make things that look far better than this. Just the other day, what looks to be a new state of the art anime video model was released. A new anime starting production today would look a whole lot different than this. And if we look forward 5 years from now, things are again going to be on an entirely different level.
So what does this mean for anime? I think the technology will slowly start to get adopted more and more as it proves itself. The early days of the anime industry was basically born out of cost cutting measures to make it cheap to produce animated content. Decades ago, we saw studios start producing 3d CG anime because it was cheaper. Most 3d CG anime still looks like crap, but you can also see the technology being integrated into traditionally animated shows and looking really nice. You can also find things these days which I would say barely even qualify as animation. Something like “The Way of the Househusband” is literally just a sequence of still images strung together. Yet we have more anime being produced now than ever before, and are also seeing some of the most beautiful anime ever.
I think we will continue to see some studios take whatever measures they can to produce something at a low cost. AI will continue to get integrated into more and more productions. It will eventually let them start making things that look cool, rather than things that look bad. And then we are still always going to have some studios that go all in and produce a really quality product, because the people involved are passionate about it.
By “embracing”, they mean two one-off project with very negative reception?
Tldr; animators are literaly dying from overwork(long hours and little sustinance) so instead of paying them more and hiring more animators they supliment the shit situation with ai ti avoid fixing it.
There is currently a labor crunch in the anime industry due to its unattractive working conditions. A 2024 report by the Nippon Anime and Film Culture Association showed that workers were overworked and underpaid, with hourly rates below the country’s minimum wage being common.
TLDR: money
Japan’s Copyright Act, amended in 2019, is largely interpreted as allowing the use of copyrighted materials to train AI tools — without the consent of the copyright holder. The law, specifically more permissive than those in the EU or the US, aims to attract AI investors to the Asian country.
It’s actually strange that Japan allows this because that country normally has very strict copyright laws compared to the EU and the United States.
Charlie Fink, former Disney producer and current adjunct professor of cinematic AI at Chapman University, feels that the use of the rapidly developing tech will “lead to a new golden age of Hollywood,” one that would be “highly democratized, because an individual could make a film for a few thousand dollars,” he told DW
If Fink is right in what he says, in the future, I think there will be a debate about whether AI is a good thing or a bad thing. Because if AI makes cinema a movement like free software and/or open source, it’s a win-win, right?
As long as the AI is also Free Software or Open Source.
sounds like japan might be desperate for the lack of anime workers.
Much of the animation takes place outside Japan these days. If you watch enough recent anime end credits, you’ll see a lot of what look like romanized Vietnamese names. And there was a scandal . . . about a year ago now? . . . when some material for an anime then in production was found on the server of a North Korean studio (probably because a Chinese studio to which the anime had been outsourced then outsourced it further without paying attention to little things like international treaties). And I don’t think the teams remaining in Japan have any shortage of recruits.
This issue, as with any business, is “can AI produce more for cheaper at an acceptable quality?” If it does make real inroads, it’ll be the outsourcing studios doing the less-important scenes that get replaced first.
The whole “they need to get permission!” thing makes no sense. When I watch an MKBHD video I don’t need to get his permission to learn from him. I don’t need his permission to learn from his style. I don’t need permission from an artist to learn from their art style, their brush stroke technique, their colour science. I just look at it, watch it, read about it, and I learn. I can then use what I learned to make new stuff and there is nothing that they can, or should be able, to do about it.
The same applies for AI. AI isn’t recreating the material that it is trained from - it’s “learning” from it. It doesn’t take the Mona Lisa as training material and then output the Mona Lisa.
It does recreate the training material. There’s literally loads of examples of it spitting out a degraded copy of an original pi CE of art with specific enough terms.
But that is already the fault of the person who gave those instructions to the AI.
Spot on.
Why […] AI […]?
- Money
“When you are in a decision that puts you in danger you accept the tool that you initially rejected.”
It seems that Japan has applied this
Meh. There really isn‘t a lot there. The short film mentioned has largely been regarded as a complete flop with no critical acclaim whatsoever. It‘s regarded so bad it’s essentially a meme. A real slop fest.
As for copyright: Using models trained on copyrighted material is more of a gray area than being explicitly allowed. They literally mention the act went active in 2019. That is before the rise of AI. So if anything they‘re just really behind the curve and not so much embracing technology.
This reminds me of the saying that Japan has been in the year 2000 for the past 40 years. Some older folks somehow still think it‘s some futuristic paradise for sci-fi nerds when that hasn‘t been the case for over 20 years.
I didn’t click the link but I’m envisioning the old school studios (if they’re still around) just using the AI to drive physical drawings.
That would be a really cool use of AI. Get the AI to do the drawings, then get things like modified 3D printers to put them on to paper/plastic/whatever medium.
It makes no sense not to. It’d be like artists not embracing using computers to animate their shows, sticking to hand drawing every single frame instead. It’d be like people not embracing cars, sticking to using horse and carriages.
AI tools can be used to streamline processes and help make things faster and better. Anyone refusing to use them out of misguided principles or morals is going to get left behind. If I can make a cartoon in 25 minutes using AI tools while it takes you 25 days to make the same thing at the same or lower quality, I’m going to win.
