Add "Contains Generative AI Content" tag to the Content Gating system
~Nanochip
With the wave of generative AI content used throughout games/platforms that run on user-generated content such as VRChat, it should be required that creators tag their avatars & worlds appropriately with "Contains Generative AI Content" when the avatar/world contains ANY generative AI assets/content.
The goal is so that users who do not wish to see such content, can filter it out through VRChat's native Content Gating system.
Screenshot below as an example.
Photo Viewer
View photos in a modal
Log In
Chdata
Hell yea brotha
SekoPuppy
This is a vital addition for the health of the platform. Without it, more and more worlds look like Roblox-tier slop, especially now that there are monetisation routes and the usual suspects are beginning to take notice.
Uzer Tekton
AI slop-generated VR environments are harmful because they create the feeling of a world without actually creating one. Many of these environments look impressive at first — huge landscapes, futuristic cities, fantasy rooms — but often feel empty. They are built from visual patterns instead of real design choices, creating strange layouts, meaningless details, and spaces that look interesting but have no purpose.
On the other hand, a good virtual world feels like someone thought about how people will move through it, what stories it tells, and what emotions it should create. In contrast, AI slop environments can feel like walking through a dream made from random images: visually impressive, but with nothing meaningful underneath.
The problem is not AI being used as a tool. The problem is flooding VR with endless, disposable spaces that replace creativity with generation. When anyone can instantly create thousands of environments without real thought behind them, carefully made worlds become harder to discover. Creators who spend months building detailed spaces and meaningful experiences must compete with a constant stream of automatically generated content.
This hurts the VR community because virtual worlds are not just images — they are places where people gather, create memories, and express themselves. If VR becomes filled with random AI-generated environments designed only to look impressive for a few seconds, it risks becoming a collection of empty backdrops instead of a place for creativity and human connection.
To protect VR creativity, platforms should limit AI slop instead of allowing it to overwhelm communities. AI-generated environments should be clearly labeled, low-effort mass uploads should be restricted, and discovery systems should prioritize projects made with real creative effort. Users should also have the option to filter out AI-generated environments if they want a more human-made experience.
AI should be labeled and can be opted out for users who prefer human-created spaces. Transparency lets people choose what they explore and understand whether a world was carefully designed by a creator or mainly generated by AI.
AI can be a powerful tool for helping creators, but it should support human creativity rather than replace it. The future of VR should not be endless machine-generated spaces, but worlds built with purpose, imagination, and human connection.
Message written with GPT and real human sentiment
M0dZ_
Uzer Tekton have you lost your ability to write? Everyone knows generative AI is bad but using AI to write your reply just makes it worse lmao
Uzer Tekton
M0dZ_ The next time you ride in a car, "have you lost your ability to walk?"
M0dZ_
Uzer Tekton at least I have a drivers license, some people use self driving cars :)
Horrible analogy btw
Uzer Tekton
M0dZ_ Next time you use google, "have you lost your ability to think?"
Maebbie
Related video from the EU Commission just dropped. Will Social VR Platforms have to comply here? https://www.youtube.com/watch?v=0Xtd8vhytuo
Uzer Tekton
This may be the key
Uzer Tekton
Sadly probably not legally applicable
> Minimal or no risk
> The AI Act does not introduce rules for AI that is deemed minimal or no risk. The vast majority of AI systems currently used in the EU fall into this category. This includes applications such as AI-enabled video games or spam filters.
However, there is clear demand for transparency, regardless of the law. It is a matter of retaining community trust and UGC health. We need AI content labelled and option to opt-out.
ElfMom
PLEASE.
While we're at it, let's add a category to the report world feature for "misleading preview" .
I am so tired of ai slop world preview icons that in no way shape or form reflect the actual world. It's one thing to make a fanciful "title card" type preview, but I have seen so many that just use an ai imageof what I assume they
wish
the map looked like, and then the map looks nothing like that.PDB
how did you find the characted use by vrchat? i need that
BinoWulf
I would even go further than this. I would just make the system more dynamic and you can choose the tags you want to remove, worlds and avatar have user generated tags for such a long time and never gets used anywhere, I wish that I could just add tags that I don't want to see.
The current world and avatar search lacks a lot on functionality...
Ryley_Oleander
i really want this so tired of ai
Kizzy
Kind of funny that a lot of the people who are against the idea of this tag and defend the use of AI in this comment section can't even be bothered to write their own thoughts. They're clearly having AI write the posts for them, the poster below me even admitted to it lol.
Heather May
What exactly is the threshold for "This (Avatar/World) Contains Generative AI Content"?
Does it apply when:
* A single AI-generated texture is used?
* AI-assisted Udon code?
* AI-assisted writing or translations?
* AI-generated meshes or textures throughout the world?
* Or only worlds that are primarily AI-generated?
Without a clearly defined threshold, this could apply to almost any modern world. Where is the line drawn? A common response is "if I can see AI, it's too much." But if the concern is generative AI itself, shouldn't that also include invisible uses like AI-assisted code, writing, or development tools?
Does first vs third party make a difference? Does the creator have to use AI themselves, or does it also count if they include a third-party prefab that contains AI-generated content?
There are popular, largely handcrafted worlds that have contained small amounts of AI-generated content for years without much controversy because it fits naturally within the world. That feels very different from "This world is almost entirely AI generated". Similarly, I was in a world from last year that also used AI content and while people commented on it, it was a small part of the world and felt appropriate in how used.
This isn't the same as age gating, though. This isn't "inappropriate to minors" - indeed, the complaint is not about minors seeing it, but about you (and others) that I assume are adults seeing it. A disclosure on the world page makes sense, but that's very different from treating it like an age-gated content warning. If you use that as a binary "do I go or not" you will be excluding some fairly well-made worlds that may contain a small amount of AI - including worlds you may not expect.
The Steam approach - a notice, but not a restriction, that visible AI is present - may be the approach closest to what you're asking, but it doesn't really answer the question of "was generative AI used". If your complaint is the rise of AI use, it's in far more than the visual arts.
AI generated thumbnails may be the best indication: If they used AI for their thumbnail, they probably used it liberally in the world as well.
* Disclaimer: AI was used in the writing of this post. I am known to ramble, and I used it to help me tighten this post up
ElfMom
Heather May
"I used ai to write this"
Opinion discarded. Write your own thoughts.
"I am known to ramble"
And so you made ai ramble instead. Congrats.
SekoPuppy
Heather May
- If the world contains anygenerative AI content, it should be flagged as such. The problem isn't the volume used, it's the fact that it's morally bankrupt from the very beginning. We should have the ability to assess content in our experiences, go "Nah, I don't want to give this any traction", and find something genuine and heartfelt instead.
- If you use AI because you ramble, then ramble! What you actually have to say is a billiontimes more valuable and welcome that some sanitised AI summary.
Heather May
SekoPuppy You understood what I said, so I was successful by using AI. By rambling, I fail to get my thoughts across. The comment was there mostly to point out that AI takes many forms, from fairly minor to pure slop. What I wrote is what I wanted to say. It wasn't "written by AI". it was written by me, but the AI helped with cleaning it up to better get the message across. The actual message is still mine and says nothing I didn't write by hand.
You still fail to address what counts as "generative AI content" though. You use the term, but don't define it. In most uses, it means "If I can see anything AI" - writing, computer-generated voice (likely not AI), images, models - and you're right. These are, save for voice, pretty much universally slop. They represent low-effort work.
But I download third party udon. I suspect it may be AI. Do I need to label that? Even if it's just code? Code Assistants are basically industry standard now - and they can take the form of anything from "smart completion" to "write a programspecs". Does that need to be labelled? Does smart autocomplete need to be labelled - something IDEs have given us since long before LLMs. Of course, I only suspect it's AI; I do have an asset I suspect may be AI but I do not use it. Do I need to label that? What about if I find out it does use AI? What about if it's only an editor-side tool? I did use an AI tool to setup texture overrides for iOS - I went to google to see if someone had written it to save myself the time (I don't work with C# normally), and it just gave me the code when I searched. Should I have turned it down and spent the 1-2 hours figuring out the API and writing it from scratch? It wasn't complicated, but still takes time to write. Should I have kept searching for a human written version instead? Took me like 10 minutes to fix a minor error, but otherwise it did what I needed - including managing edge cases)
Anyways, this is NOT written by AI. I did that to demonstrate a point. It's also focused on one very specific question: What is "generative AI" actually mean and where do you draw the line? We're very close to the point already where if you don't ignore some of it, you'll need to label everything. Even if someone writes code which isn't generative, it may well be using a library which is nowadays. What if Unity starts using AI? Well, then, indirectly AI would be used in the making of a world or avatar.
You need to draw the line somewhere.
Where do
you
draw it?Heather May
Ironically, not quite as rambly as the 2500 character limit forced me to clean it up a bunch - I was 1000 characters over to say the exact same thing. x.x
Heather May
ElfMom So you didn't read it and definitely didn't understand the point of the message. That was used to raise a point. I don't normally; I let myself ramble (though, apparently, the length limit here is pretty small so I'd have had to clean it up anyways even if it was by hand)
What do you define as "Generative AI Content", and where do you draw the line? Because it takes many forms. In most uses, it means "I do not want to encounter any AI generated content" - images, audio, voice, story, models, etc - though voice is most likely NOT actually AI. It can be done much cheaper with voice banks, but does sound less natural depending on quality of the voice bank.
There is a LOT of more subtle AI use happening behind the scenes nowadays. Stuff that doesn't replace jobs, but is still technically generative AI.
Anyways, this comment was written entirely by me, no AI used at all. I normally let myself ramble. I did that to demonstrate a point. Everything there was what I wrote, just edited by AI. The message was fully mine, nothing was added though a little was removed and the structure was reformatted to make it more coherent. But it was still all written by me and I stand by what it said, and even edited it further after the AI cleaned it up, because I didn't entirely agree with its output.
But, technically, it still used AI. Outright rejecting it says more about you than it does about me. You're willing to throw an argument out instead of considering its merits because AI was involved at all. That is, to be honest, a bit shortsighted, but it does let you keep your world untainted as much as you're in control of - there is a ton of unlabelled AI use you encounter without knowing it.
Instead of merely rejecting it, come up with a counter argument. Approach it in good faith. I will. There are many points of view on this, and plenty to be said on both sides - I have an opinion and a position, and so do you. So let's hear it.
Personally, I draw the line where it's used to put people out of work, and where it's used to replace thought with slop. But even then, we have to be careful to avoid falling into the trap of outsourcing our thinking to AI.
Load More
→