AI-generated art sparks furious backlash from Japan’s anime community | Hacker News Reader

AI-generated art sparks furious backlash from Japan’s anime community

restofworld.org

19 pointsmarban12 comments

This is missing the context of there being an active sense of disdain for artists amongst many people pushing AI. The tone of such comments range from near "learn to code" levels of dismissiveness, to snide comments about how art is too expensive, how artists are too entitled, art takes too long to make, etc. In the meantime, we see artists actually losing work to such tools already, so it's certainly understandable that they're upset, and that's not even touching on the training data issues. Events like the one in the article are an example of AI being used as a hammer and everything becoming a nail; we very much seem to be in the techno-euphoria stage of this thing.

Unfortunately for artists, I see no legal framework for them to get protection or compensation for their work. I think the prime focus right now for those whose careers are at risk needs to be evolving their style and practices to place them ahead of this technology. I also am disappointed that people have expended so much effort to automate art instead of some of the awful manual jobs there are on this planet. If the meme of the techno-utopian post-scarcity future is supposed to be a race of bored, comfortable creators and artists who have automated all required manual effort, automating the thing we're intended to be doing at that point seems counterproductive.

> active sense of disdain for artists amongst many people pushing AI

I feel as if this was spawned by many artists' disdain for so-called "tech bros." Most pro-AI people didn't seem to have any hostility before that.

> example of AI being used as a hammer and everything becoming a nail

I don't quite get it. AI image generation has its own niche.

> disappointed that people have expended so much effort to automate art instead of some of the awful manual jobs there are on this planet

Truth be told, the reason many of those jobs (I'm assuming mining, certain factory jobs, etc.), haven't been taken by robots is because the problem is much harder and more expensive than getting a diffusion model to spit out images.

Overall, this situation isn't too different from past situations. The Luddites were actually skilled craftsmen! Technology improves, certain jobs go away or decline—that's the story of the past few centuries.

I do agree that a framework needs to be created to handle the soon rapidly declining number of jobs in society.

Because, watching twitter, I see many people trying AI everywhere. One person used it to make an insensitive tribute to an artist. Another guy is badly reimagining pixelated sprites from retro games into raster renditions of the subject matter. Another is using it to process blender render frames to create foam on water.

Not all problems are solved by AI. All these things are nifty, but I wouldn't have called any of them a potential commercial success.

"To find out what if" is a good purpose on its own. Not everything needs to be practical; it just must be interesting, even for a short while. That aside, you often can't be sure of the effectiveness of a tool for a purpose without you or someone else trying to use it.

It’s hard to not have disdain for the tech bros considering how many dumpster fires have been lit using their apparatuses.

Maybe, but what even are "tech bros" anymore? The way it's being used now and recently seems to just mean "anyone who does technology-related work." It's become a vague pejorative like geek and nerd used to be. I assure you most of people complaining about AI aren't referring to Zuckerberg, Musk, or any of the traditional "tech bro" types.

> Unfortunately for artists, I see no legal framework for them to get protection or compensation for their work.

Isn't this exactly what copyright is for?

I find it sort of hilarious how for decades the importance of "intellectual property" was hammered into the public consciousness, yet all that immediately goes out the window the moment you can make money with AI and need training data.

What has happened is analogous to the Analog > Digital transition of the camera.

We are in a different era now. These programs spew out an infinite amount of garbage, so, as is the case with a camera, a human is needed to point it at a useful output.

These artists will be able to amplify their output by at least 10x if they leverage this tech, and the world will be a better place for it, because the rate of good artistic production will increase.

This rate increase may mean that the supply/demand curve gets disrupted and pay goes down, but other industries have seen the demand curve spike faster than the supply due to unforeseen market forces. If this happens, it’s a win-win for everyone.

> but other industries have seen the demand curve spike faster than the supply due to unforeseen market forces. If this happens, it’s a win-win for everyone.

What makes you confident this would happen here?

Not confident, just saying it’s not a for sure bad thing. If you 10x the number of people who can explore story space, you might capture a larger audience?

There’s a notion that AI isn’t ‘creative’; it doesn’t ‘have a soul’; that it ‘just imitates’, as if we are uniquely capable of being creative. Yet, it seems increasingly clear to many of us ‘AI bros’ that this isn’t true.

For now, I largely agree with the critics. However, I think it won’t be long before these sorts of statements could be looked back upon as short sighted or even derogatory.

> that it ‘just imitates’, as if we are uniquely capable of being creative. Yet, it seems increasingly clear to many of us ‘AI bros’ that this isn’t true.

I think to be able to adequately judge that, you'd have to chat with some artists (or at least art critics or art historians).

We know since Eliza that generated content can be quite convincing for people who are laymen in the field. However that doesn't mean that someone with more knowledge won't still immediately view it as an imitation.

Compare the discussion about GitHub Copilot, where tech people do have the necessary skill to judge the generated code. The result was that, while we were all impressed with Copilot, it was clear pretty quickly that it didn't actually think up an algorithm or design an architecture but just meaningfully combined snippets of code from GitHub.

Reply on news.ycombinator.com