I find it so... interesting that you come out of the gates swinging.

I agree there are issues not addressed, but to come out and say it's "lazy and incendiary" is disingenuous and hurtful, for the sole purpose of putting yourself on a moral pedestal above nanochip in an attempt to claim your opinions carry more weight than theirs.

Ideas are just that. Ideas. Simply because something is incomplete does not mean it is evil or objectively bad; it simply means it is incomplete. With the rise of AI, it was inevitable that such a discussion would eventually be had, and I see this as the start of that discussion. For reasons unbeknownst to me, you've chosen to squash it.

Lastly, you're operating under the assumption that such a system would be policed and that consequences would be delivered. This can be inferred from the original language, "it should be required that creators tag," in the original post. But, given the overall accusatory tone of your comment, I'm unable to extend any benefit of the doubt.

Even if this post dies, the topic will come up again. I hope next time you'll be more open, considerate, and helpful.

Yewnyx Look at how Steam did tags AI generated content and allows people to know what uses AI and filter it. There's the solution. It's up to the game developer to justify its use from there. There's tagging and filtering on sexual content there and the world didn't end, same with crypto. If someone is using AI in their work, whether for coding or content generation, they should be able to justify its use with dignity if they stand by it. There should be different granularity levels, such as AI code, AI code and Gen content, AI Gen content only, etc. And the community should be able to decided whether they want to view or partake in that content. They could be fine with AI code, but they don't want anything to do with AI art, and that's totally fair. If someone is worried about the implications of people not liking the use of AI, perhaps don't use it, if the consequence of people finding out are something that can't be handled. One could also keep their head down and turn off social media if they can't handle the scrutiny. They already chose to use AI and may not need to justify anything, just tag their work. Fielding the social implications is semi-optional, ignore or block people criticizing and keep on going, might as well.

-CMC- The analogy doesn't really hold, because Steam games are something you enjoy personally, rather than something you necessarily display. If a part of your "Blah is playing Foo" status update was "* and this game has AI content", and if you game status was completely public to anyone on the internet, someone might be able to say "get a load of this guy, playing a game that used genAI", even if you weren't particularly aware of it.

In social situations, one doesn't to actually use or want to see anything generated for someone with enough social clout, or at least loudness with persistence, to hurl accusations and abuse towards someone on account of it, just like someone who isn't a minor with a high pitched voice might get yelled at for being a minor, whether or not they've age verified, whether or not they approve of age verification or the provider for it.

Given the potential for abuse the question has to be asked whether the juice is worth the squeeze. For age verification, the answer is clearly yes, because there is a bright-line test that is objectively truthful and gets surfaced in a way that is extremely useful in adversarial situations: "is this person an adult?". Having a definitive affirmative answer is useful.

For genAI, as much as we might benefit from such a clear test, none exists, and the false negative rate is likely to be sky-high, elevating the bullying risk in a very obvious way. The problem cannot be solved in isolation from the social impact, and all solutions to moderating genAI cannot avoid the social, cultural, and policy implications that follow.

Agdistis Nice bait. No, a feature request that asks for a signal that something is immoral that does not bother to propose how to mitigate people being immoral as a direct consequence of that signal is not just incomplete, it is a witchhunt hazard. I think that being able to minimize one's exposure to pixels generated by AI is actually worthwhile, but doing so carelessly in an environment in which people easily bully each other is an act of harm and I am offended when people do rude things and avoid taking responsibility be cause the form was apparently polite. This proposal is obviously compelling enough to snowball yet too ill-considered to have any satisfying response from anyone.

YewnyxThis is important, like 18+. VRChat is given life by creators who have strong feelings about this subject, they should be empowered rather than disenfranchised. You talk as if dealing with negative social implications of AI are ‘inevitable’ in the same way that using AI is inevitable. People in this community who've come together to create solutions and come to a consensus on various matters regarding the last 10+ years of this game are going to suddenly loose their senses and turn into an overwhelming mob? VRChat has fielded community adversity before, and its our job as the reason this game continues to exist at all to scrutinize instead of accepting everything we're handed if we want to stay here happily. Most people are Grey, not B & W caricatures. There are certain things I'm totally fine with using AI code, but I'd rather filter AI art. I want the choice and agency to make that decision within this platform as I see fit.

People like Epic's Time Sweeny are upset both that users are vocally against him on social media regarding AI AND that Steam has AI tags. With UE6, he's tripling down on selling a fiefdom at the engine level to make his will of AI 'inevitable'. On the store front, Valve thankfully ignores his pathetic mewling. Consumers are happy with the labels, the world keeps turning. Speaking of abuse, people rail on Steam's review system because people use it for memes, comedy, and to spew toxic right wing nazi garbage and brigade every time a Kirkening-esque curator turns up due to GamerGate wastes of skin like Grummz saying things about women and minorities on X. The VAST majority of reviews are extremely helpful! Brigaded reviews reported by the community have been eliminated by Valve.

AI proponents who don't have the ability to buy out a captured audience should fight for what they believe in with their comrades if they think AI is worth it, and keep their heads down in the ways that I listed if fielding the social consequences is too much for them. If they want to target a skeptical audience, and it turns out they don't like the ethics involved, maybe that should've been considered. Obfuscation and normalization of a certain technology that is known to be HIGHLY unethical for ripping artists, ripping code, and ripping copyrighted material for over the last decade to the tune of Quadrillions of bytes? Give us tools to choose.

Yewnyx You fail to realize like any other content tag, this obviously isn't a black & white tag, it's entirely a gray area and should be treated as such on a case by case basis.

The content gating system exists and would be a decent starting point/foundation for self-reported AI generated content that a creator publishes onto the platform. This way, users can filter it out if they decide to.

That said, a lot of your questions assume that this is the end-all-be-all solution in an ever-changing creative debate, which I absolutely don't have the energy to get involved in.

Regardless of your view, there's no doubt it should be up to the user where they define the line on gen AI content, but at the bare minimum, this proposal is a starting-point/foundation for the self-reported gen ai content.

What happens after this? I don't know, but I'd definitely consider ideas/canny posts later down the line!

As for your jab on feeling this proposal is lazy and incendiary, I hope you have a great day and don't dwell on this too much. 😊

"I am offended when people do rude things and avoid taking responsibility"

Not sure why you think I'm trying to bait you, or with what.

I'm really not sure how you've managed to, in the same breath, acknowledge its incompleteness and still insist that VRC would push an incomplete feature.

"Doing so carelessly" - then offer some ideas or solutions. You've done nothing but kick and point fingers and exhibit the same behaviors you yet complain of.

Not sure why you're so angry, or why you think your anger will produce anything of value, but I promise you there are better ways to spend your energy.

IrbisTheSnep Absolute insane take comparing a Gen AI filter to "furries, lgbt etc" 💀

~Nanochip My sanity is not a topic of this discussion. Both LGBT and AI are topics of individual moral compasses. I think hating LGBT is bad, but its my morality compas to think that and I don't think about forcing people to like it either. Some poeple hate LGBT as strongly as AI is hated. Both topics have also supporters. This os why I would see option to give everyone an option to work with tags and give an option to make individual blacklist of tags. This way not only AI would be handled, but also other questionable content with phobias or fans, for example diapers, spiders, holes...

IrbisTheSnep Bro I'm sorry I might be tripping, but you type like an AI

FemPyro Lol. Someone in my workplace also told me something similar. Dunno if it's my age or way I learned english. I have quite scientific mind and for long time I struggled to write/share my thoughts in an easy to understand way, so I learned to use similarities to explain different points of view. I rarely use AI and mostly as a tool for initial research, to later check sources it used myself. My writing style did not change in last 10+ years, so it might be the other way, AI was trained to write like me. XD

I assure you I'm a hum... I mean furry. XD

Patroll that sounds like a rule change on how thumbnails are made then ai overall. just auto generate thumbnails from the ingame camera .

Axinovium For some reason the "a good world shouldn't need an AI-assisted warning label" and similar sentiments only ever comes from people who tell everyone that AI is inevitable and can make the end result so much better, while at the same time fighting tooth and nail against having to disclose that they're using it.

No one is owed visits to their world, least of all people who outsource the creation process to tools that only work off of stolen work.

Add the tag, and then have it stand on its own merits, rather than allowing deceptive creators to succeed off of stolen valor.

Gamekiller48 Allowing it to stand on its own merits would mean not adding tags though and people judging a world for what it is rather than pre-deciding it based on their personal ideological beliefs.

Axinovium That would only be true if AI generated content was only an aesthetical and quality question, which it isn't: it's also an ethical question.

In that way it should be treated like all other existing products with ethical issues, such as cans of tuna that have to add the disclaimer if the fish inside was caught by rod or by dragnet, or meat in the supermarket that has to declare in which conditions the animals have been kept, or labels on chocolate that allow the customer to see if their chocolate was produced on plantations where the workers received fair wages or not.

The details of production which allow the consumer to decide if the product was produced in circumstances that are reconcileable with their own ethics part of the merit by which the product needs to be judged by.

Gamekiller48thats old news , ai your talking about is not the same as world building since a ai world is not a ai picture. nore are they build the same way.

Gamekiller48 AI is only an ethical issue for people who don't understand it and have a completely incorrect interpretation of things like intellectual property rights and environmental impacts.

It's all been debunked before; it's all just journalist slop and sensationalism designed to push an ideological agenda.

Just be honest: in practice, if this filtering was added, it would simply be used as a way for AI haters to witch-hunt creators who are building art they dislike and disagree with on purely subjective ideological grounds.

Axinovium Okay dude, if you want to mindlessly parrot the talking points of AI grifters then sure, all the ethical and intellectual property questions are all solved and settled.

If you live in the real world, you might realize that A) intellectual property rights differ from country to country, and that the approach these AI companies have used (fair use interpretation) does not have legal standing in large parts of the world, and is even on shaky grounds in the US itself because it arguably violates the fairness stipulation/non-compete provision of the laws it bases itself on.

B) that ethical issues are absolutely raised by people who understand it, you are of course free to strawman those arguments away by saying that only people who don't understand the technology hold them, but you'd be wrong. Case in point: arguably a good chunk of the LessWrong community.

C) That such filters would be used for witch hunts is one interpretation, that be true, but we have no proof for that. It's also possible that it will reduce the amount of friction, as it will allow people who don't want to interact with AI generated content to simply ignore these worlds.

D) Because I know you'll just mentally earmark me as someone who "doesn't understand AI" at this point: I've developed AI systems of my own, have a bachelor's in computer science, wrote my thesis on data annotation tools, and had the lectures on ethics and property rights.

Correctly marking the contents of for what it truly is is the correct pro-consumer choice.