Filtering AI-Generated Mods

[PukupukuDragon]

We need a checkbox, toggle or personal user blocklist that lets us filter out mods in the database created using AI. I'm returning to this game after a year or two and the amount of AI-illustrated thumbnail slop that has swamped the database is appalling.

Several communities in Thunderstore have already enacted a policy where you must self-tag your mod "AI Generated" or it is removed. This solution satisfies both parties: the ones who don't want to see it and the ones who don't care. It's going to be necessary for game modding communities to have this going forward so users don't have to wade through the slop mire to find quality content. I understand Vintage Story has a smaller modding community and being as inclusive as possible is a priority but the longer a policy like this is put off the harder it will become to clean the database up later.

Please do something about this. My suggestion is to have a report button on mods that lets users flag them for review by the admins. Any mods posted/updated before the policy get the tag applied and a warning sent to the author. Any mods updated after the policy get removed.

Edited May 17 by PukupukuDragon

[Tj Pepler - Critcher]

I fully support this.  This is a multi faceted issue as well, because I don't want AI opponents to feel they have been tricked or cohearsed into using something they don't believe in.  And so I think it does save everyone all kinds of heartache, imagine if there were no AI comments at all anymore because you can simply visually skip mods you're not interested in, in that case there is no reason for a post or comment at all and moderators could safely assume any such comment was intended to be inflammatory.

I don't know about forcing anything on anyone, I really don't know what is right, but I think if people that felt very strongly were to politely request mod authors add an AI marker to their projects I bet most wouldn't mind complying. 

This is my version of voluntary disclosure, I don't think it needs to be anything more than this, if anyone feels differently though I would love to hear your suggestions.

 

[Meowsler]

Wholey support this because ai garbage just bogs down the game overall especially for anyone wanting to look through the mods database, its obvious that all the ai trash involved will either deprecate or just flood the already limited amount of good mods available. 

And yes, the idea that ai garbage should be taken down if not properly ai tagged is a perfect solution. Please add this!

[LadyWYT]

10 hours ago, PukupukuDragon said:

My suggestion is to have a report button on mods that lets users flag them for review by the admins.

Welcome to the forums! To my knowledge, there's a report feature coming to the database...for both mods and comments, as well as safeguards for false reports.

 

10 hours ago, PukupukuDragon said:

We need a checkbox, toggle or personal user blocklist that lets us filter out mods in the database created using AI.

If it would get people to stop fighting about it, sure, but sadly I think there would still be problems of users getting harrassed for AI use(or perceived use). You'd also have to set rules about what kind of AI use requires an AI tag, and then get mod authors to tag accordingly...which can get pretty messy rather fast.

So I mean...not the worst idea, but it might cause just as many problems as it solves. Probably better to just use whatever mods you want to use and ignore the ones that don't interest you(and report the ones actually breaking TOS).

9 hours ago, Tj Pepler - Critcher said:

but I think if people that felt very strongly were to politely request mod authors add an AI marker to their projects I bet most wouldn't mind complying. 

I'd like to think this, but I think that ship sailed a long time ago. At this point, the mod database is making the worst forum dramas look somewhat tame.

 

9 hours ago, Tj Pepler - Critcher said:

This is my version of voluntary disclosure, I don't think it needs to be anything more than this, if anyone feels differently though I would love to hear your suggestions.

Personally, I enjoy a short note about how generative AI was used and where in addition to the tag. Not that such an explanation is really necessary, but I find it interesting, if not helpful, to know the where and the how.

[hstone32]

Curious what people generally consider to be inflammatory when it comes to AI. I think generative content is the main thing, but the therm "AI" might refer to a lot more than just that. One of my big personal problems with generative content is that 99% of it is plagiarized; the models were trained on material without the original creator's knowledge or consent (at least if you're using anything made by OpenAI, Microsoft, or Google). I also think generative content should be used wisely, that the user should carefully consider areas where non-deterministic content is appropriate, and where it is not.

But what about anything belonging to the broader scope of machine learning? Is that also considered controversial? Let's say someone trains a ML model on animal behavior, and then uses that as a means of driving animal ai in-game. Personally, I think that is a case in which non-determinism would actually stand to benefit the gameplay experience. But would a mod that does that need to mark the proposed AI flag? It isn't generative AI, but it is still AI.

[PukupukuDragon]

1 hour ago, hstone32 said:

But what about anything belonging to the broader scope of machine learning? Is that also considered controversial? ... But would a mod that does that need to mark the proposed AI flag? It isn't generative AI, but it is still AI.

Stable Diffusion image generation, LLM-written code, machine learning evolutionary algorithms, etc all count. Absolutely anything that's ever touched a neural network within the mod itself or its mod page, no exceptions.

It's like the "no politics" or "don't be rude" policies some communities follow. Obviously some politics are going to be less controversial, like telling people to plant trees or feed the poor. And if someone is bullying, it is more socially acceptable to be rude to them to defend someone. But when you add qualifiers to rules people will bend and twist and argue about them. It's just easier to just say "tag AI content" even if that's a very broad term spanning a multitude of technologies and ethicalities.

[slimes]

I'd really like to enjoy playing this game without interacting with AI-anything. That means muting mods that use AI, muting users that use AI, and muting forum posts trying to edge people into using AI. I want nothing to do with AI while playing this game. I am trying to play this game to take a break from technofascism and I really don't appreciate debates about whether or not community safety features should be added. Please add community safety features to allow people trying to enjoy this game to do so.

Edited Friday at 05:54 AM by slimes

[Rias]

I will not name specific mods or authors as to not cause any flame wars, but I have personally experienced performance degradation & vastly increased bugs with mods that use any form of AI. Not just the ones with AI code - if someone is willing to use a theft machine to create a thumbnail instead of doing the bare minimum of an ingame screenshot + text, or hiring an actual artist (or just politely asking one!), I doubt their ability to be thorough with testing their code.

I'm very disappointed that this is even much of a discussion. GenAI is harmful for a frankly comedic amount of reasons and giving it an inch means giving it a mile. Bare minimum, all AI-involved content should be tagged as such.

[TakenMyName]

19 hours ago, Rias said:

I will not name specific mods or authors as to not cause any flame wars, but I have personally experienced performance degradation & vastly increased bugs with mods that use any form of AI. Not just the ones with AI code - if someone is willing to use a theft machine to create a thumbnail instead of doing the bare minimum of an ingame screenshot + text, or hiring an actual artist (or just politely asking one!), I doubt their ability to be thorough with testing their code.

I'm very disappointed that this is even much of a discussion. GenAI is harmful for a frankly comedic amount of reasons and giving it an inch means giving it a mile. Bare minimum, all AI-involved content should be tagged as such.

I may be biased (as an artist) but i wholeheartedly agree - if we cant completely ban it, please let me at least have a personal blocklist. Its TIRING to scroll past the same gen ai slop thumbnails on the db when looking for anything new intresting or checking for updates.

[DatDarnDoggo]

sadly it seems like, on discord at least, they are more interested in shutting down this dicussion than actually doing anything about it

[Diff]

Just now, DatDarnDoggo said:

sadly it seems like, on discord at least, they are more interested in shutting down this dicussion than actually doing anything about it

You see echoes of that on the forums. The discussions often get heated so they're quick to act on the forums.

[LadyWYT]

6 hours ago, Diff said:

The discussions often get heated

This is why many threads gets locked/deleted in general. Outside of violating the rules, the "discussion" isn't a discussion anymore, but rather personal insults and almost outright threats, in a few cases. It's fine to express displeasure with an idea by politely criticizing it, and it's fine to dislike whoever proposed/supported the idea, but what's not fine is trashing the individual for having different beliefs rather than politely debating the ideas.

[Teh Pizza Lady]

On 5/17/2026 at 12:01 PM, Tj Pepler - Critcher said:

And so I think it does save everyone all kinds of heartache, imagine if there were no AI comments at all anymore because you can simply visually skip mods you're not interested in, in that case there is no reason for a post or comment at all and moderators could safely assume any such comment was intended to be inflammatory.

Except some people can't resist making it their personal crusade. They've decided AI-assisted work is a great evil, and they'll fabricate evidence to support that conclusion, brand the code as malicious (intentional by the author or not), then push to have it removed. What drives them is simpler than they'd admit: they fear what they don't understand.

Isolating the affected mods wouldn't solve anything. There are those who would seek them out anyway, eager to weigh in with opinions that are both uninformed and unwelcome. I've watched them do it, smug and deliberate, carefully wording their objections to walk the line between "expressing displeasure" and "targeted harassment of the creator."

[LadyWYT]

3 minutes ago, Teh Pizza Lady said:

Isolating the affected mods wouldn't solve anything. There are those who would seek them out anyway, eager to weigh in with opinions that are both uninformed and unwelcome. I've watched them do it, smug and deliberate, carefully wording their objections to walk the line between "expressing displeasure" and "targeted harassment of the creator."

Maybe. But I think it could be argued that if the user had the option of filtering out the stuff they don't like, and then deliberately went and harrassed anyway...moderation actions could be more strict in those scenarios since it's harder to argue that the user wasn't intentionally looking for a fight.

[Diff]

11 minutes ago, Teh Pizza Lady said:

Except some people can't resist making it their personal crusade. They've decided AI-assisted work is a great evil, and they'll fabricate evidence to support that conclusion, brand the code as malicious (intentional by the author or not), then push to have it removed. What drives them is simpler than they'd admit: they fear what they don't understand.

Isolating the affected mods wouldn't solve anything. There are those who would seek them out anyway, eager to weigh in with opinions that are both uninformed and unwelcome. I've watched them do it, smug and deliberate, carefully wording their objections to walk the line between "expressing displeasure" and "targeted harassment of the creator."

Why avoid something largely beneficial that will decrease overall bad vibes and conflict just because it won't solve every instance? This kind of behavior is likely already covered by the rules, as evidenced by how these topics always get locked. 

Edited yesterday at 05:08 PM by Diff

[LadyWYT]

20 minutes ago, Diff said:

Why avoid something largely beneficial that will decrease overall bad vibes and conflict just because it won't solve every instance?

The better argument I can think against it, is that adding new site features tends to be a lot of work, and this kind of feature relies upon mods getting tagged correctly. Which seems like it would be rather difficult to enforce. It can't be left up to the community alone to add tags, as there's quite a lot of stuff that get accused of being AI when it's not, and leaving it entirely up to the moderators to go through everything can get expensive rather fast. And of course, resources that go into that, are resources that aren't going toward other areas of development.

Ultimately, I don't think it's the worst idea, but I do think it's better to wait and see how much of the "fire" the report feature gets under control first. In the meantime, best to support the mods you like to use, and ignore the ones you don't like--AI or human-made. Quality tends to rise to the top and lack of it tends to sink the ship rather fast.

[Reva Revenant]

Imagine filtering mods by WHO or WHAT made them instead of, I don't know, quality, uniqueness, or what they actually add to the game.
 

On 5/20/2026 at 10:37 PM, PukupukuDragon said:

Absolutely anything that's ever touched a neural network

Also can't wait for the "ethical" enemy AI that runs on ten thousand nested if-else statements because apparently using ML for NPC behavior is problematic now.

Bad mods existed long before AI was a thing.
If AI thumbnails are your quality indicator, what's your plan for the human-made mods that are just as broken and abandoned?
A "made with pure human suffering" badge doesn't really say much about the code inside.

I'm no artist and can't speak to that side of things, but in coding, AI is to Stack Overflow what the internet was to physical books.
More convinient tool doesn't make the craft disappear - it just changes where you look things up.

[Diff]

3 hours ago, Reva Revenant said:

Imagine filtering mods by WHO or WHAT made them instead of, I don't know, quality, uniqueness, or what they actually add to the game.

We do it all the time with products. Provenance is important. A lot of people loved Fairlife milk not for the milk, but for its provenance until it came out that they are actually beating those cows. Nothing physically changed about the milk before or after that news broke, or before or after they started beating their cows. Blood diamonds, "Made in the XYZ," "locally owned," Fairtrade-, Utz- and Rainforest Alliance-certified chocolates, responsibly sourced conflict minerals. It's a very, very common and routine thing to choose and identify products by provenance of their component parts. And AI does have its own ethical issues. You may ignore it, others choose not to, but it's not difficult to imagine at all. "WHO" is also an easy discriminator. If someone consistently releases poor quality mods, you may choose not to engage with any of them, rather than waste more time investigating. AI is that someone for many.

As mentioned elsewhere in previous topics, practically, AI mods have a higher failure rate for performance, for bugs, and people may choose to avoid the whole class. It's not that human-made mods are immune to having problems. But the rates are far different with 100% human-authored code vs AI autocomplete vs vibe coded nonsense. With no disclosure, it's impossible to tell how AI was used in a mod, and it's often less of a headache to just skip the mod.

Edited 1 hour ago by Diff