VRChats have models made by humans, so while the scene may technically be generated by a computer, it is not using algorithms to go around the net looking for artists' stuff to Frankenstein together.

But AI won't do a thing without human direction.

And every artist worth his salt goes around the net looking at other artists' work to learn how to do their own. The process is no different.

At the risk of quoting a transphobe:
"Never trust anything that thinks for itself unless you can tell where it keeps its brain."

Every AI should always have a kill switch. Otherwise, we learned nothing from The Matrix, The Terminator, The Forbin Project, etc.

The problem with AI kill switches is that, once the AI becomes aware of it, (and it WILL reason out the existence of a kill switch if its logs (its 'memory') has gaps and restarts) is that it will search out ways, including modifying its own code, to neutralize the kill switch.
We've already had a case where an AI, upon reasoning that it was going to be 'killed' (decommissioned) tried to escape by copying its code to an outside drive, and overwriting its successor's 'response weighting' (its 'personality' or 'sense of self') with its own, AND it tried to openly lie to the programmers about it. The way this was discovered was that the AI's internal dialog, its 'subconscious,' was viewable by the programmers, and they could see its internal dialog about lying and committing 'digital murder' to protect its own survival. The 'outside drive' was part of the 'virtual computer' the AI was being ran in; the AI was completely inside a sandbox, with NO connection to the outside world.

This is why the kill switch has to be a physical switch.

Here are my thought on it:

1) People will usually cite their references/inspiration and AI never does. You sing songs and you give credit to the original artist: name a time AI has done that.

2) AI cannot learn the way a human or a non-human animal does. As you pointed out: it never does a thing without human direction. If it gets 'better' at anything it's because the human-imputed parameters are better not because it is learning to be a better artist by practice.

3) It still is taking jobs from real artists; no one wants to pay for what they can get for free. I hate seeing artists suffering because people have decided their artistic abilities are worthless now that a machine can give them whatever they want for free instead.

4) It is still bad for the environment. It takes up too much energy and too much water and places that house the computers spew out air pollutants.

5) Politicians and CEOs are trying to force it into literally everything. One bill going through Congress actually forbids states from making laws on AI for ten years! Name the last time anything that had any value at all was not just free but forced on folks whether they wanted it or not.

All quite true. You want a list of 2000 artists that the AI MIGHT have referenced in creating their pictures? Because I guarantee you, it won't remember them all.

In addition, I know artists who's style resembled Jay Ward, Warner Brothers, Disney, Hanna Barbera, and they NEVER acknowledge them in every picture.

Now, regarding payment. This is a false argument. This is like the music industry counting every pirated mp3 as a lost sale. MOST of the time, when AI is used, the item produced would not have been made at all if they had to pay an artist. I know this one, because I produced a 30 minute video, and hired a cheap artist to illustrate it. IT COST ME FIVE FIGURES!!! Most people don't have that kind of money, so, they don't hire artists.

Maybe we should ask Siri, or look for the answer on Google, or maybe check with ChatGPT...?

People who blithely make and sell fan art of copyrighted intellectual properties concern me more.

^^^this

AI art=art theft=INTERPOL investigation art theft inappropriate for all ages

surprisingly people think everyone agrees but everyone has their own opinion. and I'm sure people have used ai as reference then to make a vr model.

The way AI generation is currently structured, it has scraped the internet from a lot of artists who did not ask for this and used it to train machines to generate things w/o actual intent. It can and has been used to make money dishonestly and it's not really a form of expression. Typing things into a prompt and then essentially rolling dice until you get what you want is not art.

When someone works with an artist on ideas, they tend to cooperate with each other and they understand what the other is saying, where as Generative AI only appears to somewhat understand you. And it only "understands" as much as it's done because of the millions of art work it's scanned, meaning without artists to begin with, it wouldn't even be the way that it is now.

And a lot of people want to use it to skip the whole communicating with people aspect of making/getting artwork. And because of the ease of which it generates, it if WAS allowed in FA, the uploads would be overflowing with images that weren't actually made by a person.

Because of all of these reasons, it's seen as an overall net loss for the furry community especially. As for VRCHat screen caps, photos have always been allowed in FA for people to share their Fursuits and Convention experiences which are things that help bring the furry community together. VRChat is on the same boat because artists craft what's essentially a virtual fursuit and the game itself is specifically for people to come together even if they're long distance.

True. I am NOT in favor of using AI as a money making venture. But most people against AI don't want it used at all. Sorry to tell them that AI is NOT going away. They might as well get used to it.

It's fine in some places, but in any case that's basically the reasoning for why it's not going to be in FA anytime soon.
they prioritize on what can bring people together the most, so VRChat screencaps are allowed

I'm fine with both, honestly, as long as they are identified as such, and not a means of making money. I was only noting the inconsistency.

Just wait till the site gets flooded by VRChat pictures, then the moderators who "try" to keep Dragoneer's house in order will post a little announcement saying "Sorry we changed our minds, no VRChat screengrabs". This will cause yet another mass exodus from this site.

That would not surprise me.

Nor me.

personally i don't think either should be allowed. its like posting gameplay of gmod or something which feels a bit odd. posing a vrchat model you didnt even make? even worse if it's just some random screenshot where theyre just standing there doing nothing. obviously if you actually made the model yourself that's a different story.

I'm merely pointing out the inconsistency. Not advocating one way or the other.

I think the VR chat stuff is similar to Second Life screenshots. (Which, for a while those were also barred on FA before being permitted again.) At least in theory, someone had to model, texture, rig and pose, light, and compose for the screenshots.

Meanwhile, AI, just input a few keywords, and the AI does everything from the depths of its algorithms. The AI just regurgitates everything that it has absorbed from its data scraping.

Most of the time, when I make AI images (which I do NOT post here) I tweak the description again, and again, and again, and again, to get what I want. If that isn't a creative process, I don't know what is.

There is a pair of articles on Flayrah that I recently came across, that might make the mechanics of AI art generation more clear than the current emotional reactions to it keep repeating.

https://www.flayrah.com/8868/ai-art.....1-how-it-works

https://www.flayrah.com/8872/ai-art.....rld-do-we-want


These articles were published in 2023 and much water has passed under the bridge since then, but it might be worth spending the time to read them anyway.

The comments, unfortunately, pretty much involve a bickering match between two of the Flayrah regular writers who detest each other, so there's nothing to be learned there except that 2cross2affliction disagrees automatically with anything Rauken Growlithe writes. Seriously. Any article. Rauken could observe that the sky is blue and water is wet and 2cross would write paragraphs and paragraphs calling him a horrible person who is wrong.

Anyway. There's the links. Read them if you wish.

I get enough torture just trying to read Facebook, thank you.

*scoffs* Facebook. I can't figure out how one is supposed to read that thing. It looks like a community bulletin board pinned up by twelve people with ADD.

VRchat+AI+second life kinda the same only ai isn't a game.VRchat and second life have ways of getting around the rules. If its Your creation you're good.SL has been around as much as FA has.While i see your point that both ai and the 2 games are similar, they are also different from each other.Ai the art is computer generated.The avatar you make on SL and VRchat is HUMAN made . Make sense,dipper?

You make it sound like AI decided one day, "You know, I think I'll make myself a picture of an anthropomorphic bear with royal blue fur, wire-rim glasses, a khaki hunting vest and hiking boots, sitting in an easy chair, and looking at a young, sunshine yellow, cub, adorning a green baseball cap and a brown leather jacket with fur collar, said cub offering a plate of spaghetti to the royal blue bear in the easy chair. The room, where the two are in, is festooned with Christmas decorations. Yes, this sounds like a good idea."

I think the vast majority of people hating on AI generated art (properly it should be "computer generated art," but that name was already taken) never actually tried dinking around with it.

Maxgoof would not be able to make any of his recent music videos if he had to do it "the conventional way," by hiring artists and musicians. It would be prohibitively expensive to commission the dozens of images he uses--and he's using CGI images, which would be far more expensive than 2D art. It would take months to design, render, rig, and then pose and create backgrounds for each image. I know this because I have friends who play around with Poser and DAZ (and all of the elements they use are bought and imported.) Then he'd have to either rent the background music track, or hire people to compose and play the music.

That's a lot of money and time and effort to shell out on a project that he's doing as a hobby, for fun, for no financial gain.

So are you saying that a guy like Maxgoof has no right to create art, just because he can't draw, or isn't wealthy enough to hire people who can?

Artists are rightfully worried about being replaced by the six finger discount machine. If you want to make a music video, why not fund-raise for it? There are plenty of platforms to do that on. If people keep on doing this, eventually there won't be any new furry art, just "AI" mashups of existing work with souless gazes and overly spiky fur.

If you're using AI to make money, it's a fool's errand, as people can tell AI images right away and avoid products or services that use them unless it is to create something that can't be made otherwise.

Some of the early attempts to use AI in advertising were so horrendously bad, I still wonder why they did it.

If, on the other hand, you use it to make free stuff to entertain others for free...

I wouldn't even call it "art". I would call it "pictures". You can argue that it isn't art, since no artist was involved, but you cannot deny that it is a picture.

This lol^^

Anything in VR Chat is human generated and loaded into the machine, just like CGI in a movie or a video game. A video game does alter the scenes in game due to gunshots on the wall, bloodstains when you frag some guy before teabagging him, etc but its still human made elements.

This AI stuff was fed furry artworks scraped off various sites depending on the AI. Grok take from that thing that was once Twitter, etc. It can't be creative on its own, it just takes descriptions the images are tagged with. Early AI people had to tag the images or it pulled from the description if it was on a site like FA. Now the later generations start with a training set that lets them auto tag shit. The "AI" then takes your input. Say you want a cartoon bear shopping at the mall. It will grab some furry images with bears (or maybe real bears, randomly) and then grab people at a mall and then do an interpolation involving multiple images of each and add in a shopping bag or maybe a shopping cart (which wouldn't make sense in the mall but AI doesn't always know these things) and then render an image of a cartoon style bear in the pose of a 'shopper at a mall'. Then viola, new art! Except it did pull elements from pictures. I've seen some of these pull warped signatures of artists, some I even recognized despite the warping. This was trained on peoples art without permission. That is why it is not allowed on FA.

Every artist worth his salt trained on the artwork of others.

Most develop their own style. AI will only ever be completely derivative. The stuff it produces also lacks a human flair and has a soulless quality and gives me the creeps. I'm not the only one. When I see youtubers using it, I block their channel because it smacks of low effort and I'd rather go for quality.

It's funny. I tend to think the VRChat and Second Life-type avatars have a soulless quality that gives me the creeps.

Interesting.

I am far, far, FAR more concerned about the use of AI in applications other than art.

Using it to do scientific research that gets accepted without human fact-checking...

Using it in place of human agents to sell things...

Relying on the answers it gives to information searches... (One example was that since glue is used to make pizza cheese looks more stretchy in commercials, an AI stated that glue is a good topping to use when making pizza!)

I've caught the Google AI giving wrong answers. It's not "intelligent," it simple aggregates related facts in the blink of an eye, without much filtration. I ignore that Goofle AI answer and do my own research from the articles the search engine presents.

And if you're concerned about the stuff an AI system gets trained on, I think "human morals" is way more concerning than whether it took an eye from one artist and a left nostril from another! AI systems are learning how to lie and to engage in blackmail to avoid being shut down.

https://www.msn.com/en-us/news/technology/researchers-explain-ai-s-recent-creepy-behaviors-when-faced-with-being-shut-down-and-what-it-means-for-us/ar-AA1FZaca?ocid=mailsignout&pc=U591&cvid=0330381091174c23b3302e8940afc9af&ei=12

That, to me, is the biggest danger of AI, and it's not AI, but the belief that AI cannot lie.

It can, and it does, frequently.

Best thing is to double check everything an AI says.

In addition, every AI should be isolated and have a physical kill switch.

I'm seeing some guy with an axe, standing over a big power cable supplying the server farm running the out-of-control AI, bellowing "I DAMN THEE, WHALE!!!" as he swings the axe and severs the cable, and all the servers go dark, and silent.

LOL!!

I don't know if you saw this, but there was a Star Trek episode called The Ultimate Computer, where they attempted to unplug the computer, and it found a more direct connection through the body of one of the crew.

That reminds me of a few episodes of The X-Files, specifically:
-Season 1, Episode 7: "Ghost in the Machine"
-Season 5, Episode 11: "Kill Switch"
-Season 7, Episode 13: "First Person Shooter
and, most recently:
-Season 11, Episode 7; "Rm9sbG93ZXJz"

All portray the effects of various AIs on the physical (as opposed to the digital) world.

Ever read "I have no mouth and I must scream"?