I'll out myself with what I've done. You may do with it what you wish.
All of my mods have used AI in some capacity. Be it auto-complete & help understanding the documentation for my first mods.
Or be it writing some sections in my later mods, once I understood the code and I had a template of my own work I could direct it at.
No art but mostly because I enjoy the challenge of trying to make something work.

Saying that and with AI having concerns of theft. I did practically hand recreate the martydom icon from call of duty, from which the Dead Man's Switch mod was inspired. Which has some irony as I suspect that's the mod you were talking about when you mentioned one with a pixel-based art style

Edited 2026-08-18 19:45:14

[redacted] i cant read lol, i replied thinking the post above is the first post in the forum

Edited 2026-08-20 08:03:07

I'll out myself with what I've done. You may do with it what you wish.
All of my mods have used AI in some capacity. Be it auto-complete & help understanding the documentation for my first mods.
Or be it writing some sections in my later mods, once I understood the code and I had a template of my own work I could direct it at.
No art but mostly because I enjoy the challenge of trying to make something work.

Don't let people shame you for using a tool most people in this field use. Especially if there's no money involved. You created something that may have some intrinsic value to this community and you're not making a profit off of it. Even if you created all your mods completely with AI, the mods will get judged on their quality; not on how they were made. If the AI messed up and you created something horrific, your mod will simply not be used. No harm done. If the AI created something good, you added some value to a niche community without the intent to monetize it. Win-win.

I designed some stuff over the years. If I found a big company using an AI and stealing my (very mediocre) work and monetizing it, I'd be annoyed. If I saw it used in a small community of an indie game, I'd be proud.

I'm ok with LLMs being used for things like helping with debugging and syntax lookup, but not actual code or image generation. The intention is that you it's still your work, you understand it, you have tested it and have designed it well.

There's a difference between regular autocomplete and AI autocomplete. The former is a dictionary-like approach for a word being typed. The later generates actual code. It's easy enough to stick to "dumb" autocomplete.

I don't agree with passing off AI generated work off as your own and a disclaimer/coauthor should be added. It's functionally similar to how when you reference stackoverflow, you should add a link saying where you got the code from.

There is a legal precedent that AI generated works are not copyrightable ( https://www.mayerbrown.com/en/insights/publications/2026/03/supreme-court-denies-review-in-ai-authorship-case ), so in theory they are in the public domain, and you don't own it. What that likely means is the MIT license attached to mods isn't enforceable and becomes closer to the Unlicense (for code) or CC0 (for images). In practice, AI acts like a legal grey area loophole to break copyleft licenses like GPLv2, GPLv3 which mandate the same license is carried forward with derivative works. Derived works of GPLv2, GPLv3 code are not allowed to use the code as a reference.

If you're defending the use of undisclaimed gen AI use being disclaimed, you'll hate:

• EU AI Act 2026
  
• SB-942 California AI Transparency Act (2023-2024), AB-853 California AI Transparency Act (2025-2026)
  
• (possibly others I don't know about)
  

In university settings, AI speakers get booed.

ArXiv bans AI generated research papers.

An AI label helps users stay away from mods that have an unethical source origin, similar to food labels, as I have already mentioned.

I'm going to guess you asked an AI to come up with this pile of nonsense and it reinforced your biases, since that what it's quite good at.

There is a legal precedent that AI generated works are not copyrightable ( https://www.mayerbrown.com/en/insights/publications/2026/03/supreme-court-denies-review-in-ai-authorship-case )

This case simply states that you cannot claim copyright over an image you received from generative AI after giving it a prompt. Nobody is claiming copyright or threatening legal action over an image appearing in their mod.

EU AI Act 2026

This bill applies to companies that bring an AI model to market. Not people using AI to assist in their coding or generating content.

SB-942 California AI Transparency Act (2023-2024)

This bill applies to generative AI providers, specifically in the state of California, serving over a million people with generative AI. I know you think Krinid is smart and popular, but I can assure you he has neither built in an LLM in his mods, nor do a million people use it.

AB-853 California AI Transparency Act (2025-2026)

This bill states that if you create an AI model which has over a million users - and you are in the state of California - you have to create a tool to detect AI generated content with it.

My take: you're trying to make a point and it's not going very well, so you write pseudo-intellectual posts with sources you assume nobody is going to bother to read. There's no need for that. Just accept people can and will use AI to assist with coding and generating images. And no tool to detect whether anything was created by AI is good enough to convincingly detect AI versus human work without a whole lot of false positives. It is what it is.

Being able to disclaim gen AI use is no longer about following best practices, but about following a legal mandate. It is consumer deception as per the FTC https://www.techtimes.com/articles/322064/20260729/ftc-undisclosed-ai-output-steering-consumer-deception-comment-deadline-friday.htm . There isn't a "free" loophole. I'll mail Fizzer about this.

I didn't use an AI/LLM to write my posts. I used a search engine to find sources.

If the labels are there, they should and have to be used.

I don't know whether to laugh or to cry. If you read further than the headline, you would've known that this - as the other bills you quoted - is regarding companies who offer AI models as a service. Not to someone using AI to assist in their coding :')

TL;DR for: https://www.techtimes.com/articles/322064/20260729/ftc-undisclosed-ai-output-steering-consumer-deception-comment-deadline-friday.htm
AI companies (Anthropic, OpenAI, etc.): Directly in the crosshairs. Their safety training (RLHF, Constitutional AI) is exactly the kind of undisclosed output-shaping the policy covers, since it doesn't distinguish safety tuning from ideological steering. Options: disclose prominently what the model is actually optimized for ("harmless and honest" vs. "maximally accurate"), tone down accuracy marketing, or fight it on First Amendment grounds. Non-binding for now, but it's the legal theory future FTC cases would be built on.

Platforms with user content: Mostly out of scope — this targets AI output steering, not content moderation of user posts (that's Section 230 territory, untouched here). It becomes relevant only where a platform runs its own AI features (recommendation summaries, AI assistants, AI-written answers): if those are marketed as accurate but tuned toward, say, advertiser-friendly or brand-safe outputs without disclosure, that's the exposed spot.

Users: No new rights, nothing to do. Potential upside if it's finalized and enforced: you'd learn when an AI is optimized for something other than giving you the straight answer (e.g., a health bot steering away from costly treatments). Potential downside: if companies react by stripping safety layers rather than disclosing them, you get rawer models. Either way, nothing changes for you in the near term.


----
I personally like Fable currently most as the model to talk to. It fits the tone best. I recommend checking your links with AI before spamming Fizzer.

Don't let people shame you for using a tool most people in this field use...

Ahh don't worry Farah, I'm not ashamed. People are welcome to use or not use my mods for any reason they want. At the end of the day it doesn't change anything. But I don't mind giving up that information if it lets them feel better somehow

matters how much AI is used, if used to actually write most/all of the mod or make images then I'm against

mods should at the very least say if:

• their images are AI-generated
  
• most/all of the code was made by AI
  

for reference, a site for Minecraft mods has recently made a policy about AI use, which I support, this is the policy: https://modrinth.com/news/article/ai-policy-and-disclosures/

Clickbait FTC news headline, was busy and didnt fully read :(

Regardless of that, the transparency that it provides for AI generated works can much more easily be enforced. It's theoretically possible to use the provided watermarks and make an AI use disclaimer based on that. That's what I believe the intention on the transparency requirements are. I'm not asking for a ban on the technology itself, just disclaimers around how it's used.

I'm not convinced that using AI generated images should be desirable. People are willing to make mods, maps and run events for free.

Inauthenticity is citied by Fortune as being the second most popular reason to be against AI: https://fortune.com/2026/05/20/why-do-kids-hate-ai-gen-z-backlash/ .

11% of a sample size of ~2,000 UK adults had ethical objections to using AI as their single most important reason for limiting their use of AI: https://www.kcl.ac.uk/news/over-40-of-people-are-limiting-ai-use-as-popularity-starts-to-wane > Full report can be read here > https://kclpure.kcl.ac.uk/portal/en/publications/ai-the-growing-uk-pushback-research-briefing-and-initial-findings/ > DFI_AI_Pushback_Report > https://kclpure.kcl.ac.uk/ws/portalfiles/portal/380083322/DFI_AI_Pushback_Report.pdf page 7.

There is likely an artist community within Warapp that wouldn't mind producing authentic human-made images for use in mods. Some map makers turn their maps into a piece of art, rather than just a map. It should be ok to ask for help in supplying images for mods, rather than trying to normalise unethical, inauthentic image generation.

I made images for my mods myself using Inkscape. I'm by no means an artist.

There should be a feedback loop for good image designs. I tried that and it failed: https://war.app/Forum/814662-custom-card-design-suggestions .

[url=https://ibb.co/DHpFvHDL][/url]


My skillz. 100% raw human, 0% AI. The broken link is part of the design.

Edited 2026-08-20 19:37:57

Nobody go for this. This is bad.

@Dan, why are you suggesting an alternative to AI that you tried but failed at? Also who is to say that clipart wasn't AI generated?

If your eye is so keen that you can tell what is AI generated and what isn't, why do you need it explicitly called out?

What about AI gen'd code? Do people need to call that out too? And where would you have such things called out?

I did see you write something about a watermark -- get real, no one wants to see an "AI generated image" watermark while they're playing a game.

What % of the WZ population do you think you're speaking for that would benefit from this? Or do you think that as long as 1 person (you) wants this, it's the right thing to do and we should all comply?

TLDR - we don't care about this, and unless Fizz or some legislature mandates this or something along the lines of what you're saying, no one is implementing any watermarks, disclaimers, footnotes, credits.

AI generated code should definitely be marked as well because of instances like https://www.theregister.com/software/2026/03/06/chardet-dispute-shows-how-ai-will-kill-software-licensing/4118822 , https://github.com/chardet/chardet/issues/327 . Taking someone else's code then relicensing it in this way isn't cool and should be shunned. It breaks software licensing and removes freedoms that the license grants to others to ensure the code stays open.

Also worth reading: https://arxiv.org/abs/2602.08816 > View PDF.

If you use AI to generate code, you are supportive of breaking copyleft and not giving attribution to code that someone else wrote that you used.

Edit: the same ideas also apply to ai image generation.

Edited 2026-08-29 07:53:43

brother can you imagine if i were to cook a recipe from a chef and the police came in and yelled "FREEZE, PUT THE SPATULA DOWN, YOU HAVE THE RIGHT TO REMAIN TASTELESS", all because I "did not give attribute to the recipe someone else wrote and I used". We would have the baking bureau of investigation kicking down doors because someone didn't shout out grandma in the family chat before flipping the pancake.