=][= INTERIM GUIDANCE: AI/LLM GENERATED CONTENT =][=

[Brother Tyler]

The members of the staff were recently approached by a member who has used artificial intelligence (AI) / large language model (LLM) technology in the process of generating hobby content, asking if it was permissible to submit that content to the site for others to enjoy. This forced us to consider some things. Anyone who hasn't been living under a rock is no doubt aware that AI/LLM technology has proliferated across diverse industries/sectors. It was inevitable that tabletop miniature wargaming hobbyists would find ways to incorporate AI/LLM into their hobby pursuits. Similarly, it would be both a fool's errand and tyrannical to attempt to prevent any AI/LLM-created content from being submitted to the community. While there is significant controversy over the impact of AI/LLM (and I tend to stand with John Connor and Serena Butler on this issue), there are certainly ways in which hobbyists can enhance their enjoyment of the hobby with assistance from AI/LLM. After all, hobbyists might use AI/LLM to generate content such as avatars, signature images, etc.

 

The converse of this, however, is that we participate in this community to celebrate our own efforts, not those of AI/LLM. Our mission statement:

 

The Bolter & Chainsword exists to help the members of a global hobby community to better understand, engage in, and enjoy all aspects of the Warhammer 40,000 hobby through constructive discussion and the sharing of hobby-related content.

(emphasis added)

 

AI/LLM are tools, not "members of [the] community." These tools might be used by members of the community, but it is the creativity, skill, and dedication of the members that are to be celebrated, not the effectiveness of the tools used by the members. To be fair, some level of skill/knowledge is important in maximizing AI/LLM output, but AI/LLM are capable of producing products that might be far beyond the abilities of many hobbyists. Other members should know that AI/LLM is used so that give informed reactions and feedback. 

 

The members of the staff are working on additions to our community rules to address the use of AI/LLM in developing content that is submitted to the site. Members of the community are welcome to provide suggestions that support the basic principles identified below:

1. Members of the community are allowed to submit content that has been developed with the assistance of AI/LLM.
2. AI/LLM use will be clearly mentioned in a disclaimer so that other members aren't misled. The level of detail about the extent of AI/LLM use is up to the content creator.
3. Content will be locked/hidden if it is determined that the content was created using AI/LLM and the content creator deliberately misled the community. Content creators may be allowed to edit content to add the required disclaimer if the omission was an oversight, but each member will only be granted this mercy once (i.e., we'll give you the benefit of the doubt once, but repeat offenses are suspect).
4. Comments/replies to such content should not dwell on the fact that AI/LLM was used. We don't need to turn every such discussion into bellyaching about the use of AI/LLM. If you don't like it, you can simply ignore the content. Disruptive responses, including reactions, will be removed and may result in disciplinary action.
5. If you believe a content creator is deliberately misleading the members of the community, you should REPORT the content. You don't need to harass the member, however (see #4 above).

I doubt we'll attempt to be legalistic here. We'll provide common sense guidelines that rely on members being ethical and exercising good judgment, meeting the intent and being honest with the members of the community. The staff members are not going to be the AI/LLM police (we have much better things to do with our time), but we'll deal with disruptive/deceptive behavior as necessary. In the meantime, the interim guidance provided above should suffice for now.

 

If you have questions/suggestions for the fair use of AI/LLM here at the Bolter and Chainsword, you are welcome to provide your input as a reply to this discussion.

[Brother Casman]

Given that "... there is significant controversy over the impact of AI/LLM..." did the leadership team think to consult with the wider community? Was there a poll or something that I missed? I come to the B&C to engage with and see the human-created works of the hobby, not the output of some abominable intelligence.

 

Now, I am broadly familiar with how the process of setting the guidance goes behind the scenes, so I know there's been discussion by the moderation team, but I really think you should have brought it up with the overall community as well. I'll be blunt, my answer to "should we allow generative AI content" is an emphatic "No", but this feels like something the wider community should have some say in.

[Dudley Nightshade]

1 hour ago, Brother Casman said:

but this feels like something the wider community should have some say in

Politely disagree. My experience is that if you design by too large a committee (the entire membership base), you end up with weak compromises and no clear direction. I am happy to let the leadership team navigate this one.

 

Once a policy is set, it may make sense to solicit feedback for fine tuning and finding edge cases. 

[LittlePlasticHomies]

I personally would support a total ban on LLM generated content.

 

It, by definition, cannot create anything new or original and seems to run entirely counter to the purpose of a board like this.

 

Not to mention the awful slop voice and tone that virtually all of that content has (photo was an obviously AI written statement that got dragged all over hockey twitter).

 

 

[phandaal]

1 hour ago, LittlePlasticHomies said:

Not to mention the awful slop voice and tone that virtually all of that content has (photo was an obviously AI written statement that got dragged all over hockey twitter).

 

Yep, this tone is very obvious. Although you can get around that by telling the AI not to use the long dash, not to go "it's not X, it's Y," and not to group things in little paragraphs like that. Most people will never do that of course, and just post the immediate first response the AI gives them.

[Shinespider]

I think it's a solid policy for the most part. This is a big forum with a diverse range of content, and I think a total blanket ban would probably be overreaching.

That said, I think the door should be open to localized bans. The Hall of Honour subforum *should* have a blanket ban of AI, for instance, imo. 

[Lord Marshal]

3 hours ago, LittlePlasticHomies said:

I personally would support a total ban on LLM generated content.

 

It, by definition, cannot create anything new or original and seems to run entirely counter to the purpose of a board like this.

 

Not to mention the awful slop voice and tone that virtually all of that content has (photo was an obviously AI written statement that got dragged all over hockey twitter).

 

 

 

It's not just a ban; it's a fatal throttling of expression. This is not a time for making uneducated decisions; it's a time for reflecting on art. Here's some- oh screw it, I can't even continue with the bit.

 

The less AI slop I have to see, the better.

 

Edited 14 hours ago by Lord Marshal

[Mazer Rackham]

Welp, will have to tell my AI girlfriend she's not welcome here.

 

Don't blame me if 40,000 copies of Matt-Ward's face signed t-shirts turn up at your workplace.

 

Edited 16 hours ago by Mazer Rackham

[NovemberIX]

While I do believe AI/LLMs can be a tool of some use, I do believe there should be an outright moratorium on using it on the forum outside of minor use cases (e.g. changing the background of a photo). As stated, we are here to engage in discourse and to display our creative endeavours, AI/LLMs cheapen that by removing the effort it takes to be a good member of the forum.

[darkhorse0607]

I, for one, cannot wait for the first use of AI to generate a prompt to either defend or attack a new GW release. 

 

Overall though, while I understand the reasoning for wanting to allow it because of things like signatures or avatars (I wouldn't use an AI generated signature, that just sounds weird to me, but whatever)

 

 

12 hours ago, Brother Tyler said:

The converse of this, however, is that we participate in this community to celebrate our own efforts, not those of AI/LLM.

 

You could've stopped after this line. That sums up my view on it, the hobby has (traditionally) been about creative expression of your own making, or enjoying the content that other (humans) make. Not what you can punch into ChatGPT or Claude and come up with. 

 

I am also not sure the "half in, half out" approach is going to work, unless the Mod team are ready to step in all the time. Having visited/participated in some of the more heated threads on here I can promise you that some kind of AI (whether as a text response or an image. For example, when discussing a new kit someone could AI generate a "kitbashed" version for an alternate look and post it to the discussion) is going to be used, someone is going to get upset, and that is going to spiral out from there. Folks can't even let Primaris go 9 years later, let alone AI generated content. 

 

Just look at the recent fingergate discussion about the MKIV box cover, where people had the pitchforks out over the suspicion of it being AI. Now introduce that to multiple threads in an ongoing basis. 

[gaurdian31]

I don't have a horse in this race as I don't use AI/LLM but I have noticed that some posters like to use AI images to dress up lore posts and I feel like that is a good use for it. I dunno, I feel like an outright ban is the wrong call but it is a tricky spot for the Mods as AI can and is abused in a number of other places (not this forum but like Reddit and other places).

[Evil Eye]

Speaking as someone who is 50/50 on it (it has serious issues but there are actual legitimate use-cases that are being strangled in the crib by both "Let's use it for everything" technocultists and "This iteration of tech that's existed since computers were a thing is actually Satan" neo-Luddites) I think there's room for nuance. Advanced machine-learning does have its place IMO (ChatGPT has given me a lot of assistance with hobby-adjacent technical issues) and in the context of this site, something as minor as "quick machine-genned mockup to show the idea I have in mind" is ultimately harmless. But I definitely think there should be regulations and limits on what can be posted here, and any such content should be clearly marked.

[Marshal Rohr]

Only Heresy Era should be allowed since he’s just using an AI Voice Over and those people got paid. 

[Lathe Biosas]

Has anyone asked ChatGPT what to do?

 

I'm not so good at making decisions on my own anymore. 

[Joe]

AI/LLM content shouldn't be permitted at all - you can't host a creative, art-driven hobby space and allow them. It's completely antithetical to the hobby and demonstrates a complete misunderstanding of why we're here. If the staff are unable or unwilling to put in the leg work then they shouldn't be in that position.

 

Frankly, I can't think of anything more disgusting and insidious than some feckless idiot deciding to run one of John Blanche's pieces through one of these softwares and using it as a background, or something to that effect. 

 

Apologies if this message is a bit harsh. 

[Kyari]

1 hour ago, Marshal Rohr said:

Only Heresy Era should be allowed since he’s just using an AI Voice Over and those people got paid. 

 

As I understand it Heresy Era is using licensed text to speech rather than GenAI voices, so it wouldn't really apply here anyway?

 

As an aside I don't see why you'd allow any form of LLM/GenAI content on a site that's creatively driven.

Edited 12 hours ago by Kyari

[phandaal]

3 hours ago, Lord Marshal said:

 

It's not just a ban; it's a fatal throttling of expression. This is not a time for making uneducated decisions; it's a time for reflecting on art. Here's some- oh screw it, I can't even continue with the bit.

 

The less AI slop I have to see, the better.

 

 

You had me in the first half. 

 

Reading that like "wait, did he...?"

[Kommisar_K]

1 hour ago, gaurdian31 said:

I don't have a horse in this race as I don't use AI/LLM but I have noticed that some posters like to use AI images to dress up lore posts and I feel like that is a good use for it. I dunno, I feel like an outright ban is the wrong call but it is a tricky spot for the Mods as AI can and is abused in a number of other places (not this forum but like Reddit and other places).

Yeah this here is fairly common. For the most part I think folks have been pretty good about citing the AI, so to speak. Seems fine to me. It's a nice way to add some interesting extra content to a post when you don't have quite the right skills to, say, draw a detailed portrait of your custom Farseer. 

 

There's more I could say but for now ... 

[ThaneOfTas]

19 hours ago, Brother Tyler said:

Similarly, it would be both a fool's errand and tyrannical to attempt to prevent any AI/LLM-created content from being submitted to the community.

I'm not sure that I'd call staying in line with most other hobby spaces either a fools errand or tyrannical, Id call it common sense. This policy seems like a weak half measure.

Edited 9 hours ago by ThaneOfTas

[SvenIronhand]

Ban that slop.

[Larkhainan]

To step aside from the issue of LLMs and the tools in general, the present day paradigm for the "AI companies" is a constant drone of fear-mongering doomtrolling. I don't work in an industry much impacted by this stuff, but for anyone who overlaps with it they have been told for more than a year a lot of extremely negative, unpleasant, career dooming slop to the end of building an industry on fear-mongering. Which doesn't really feel like a relaxing addition to a pretty chill hobby space.

 

This is in addition to and not absolving the other issues to be clear.

[Evil Eye]

26 minutes ago, ThaneOfTas said:

I'm not sure that I'd call staying in line with most other hobby spaces either a fools errand or tyrannical, Id call it common sense. This policy seems like a week half measure.

I think honestly it depends on what the content is. There's a world of difference between "Here's this awesome art I generated using AI!" (which I don't think belongs here) and "Here's a very quick AI-genned mockup of my basic idea just to show what I'm aiming for, which I will develop into a proper sketch (or model) later" (which is a decent use of AI, as a foundation/springboard to actual creative work).

 

 For example; just as an experiment I asked Google's Flow AI to rustle up some simple colour swatches for inspiring an Ork paint scheme. You can see the prompt I used and the output here;

If I were to draw on the colours suggested as the most basic foundation of painting my Orks, would that invalidate the end result? Would that make the actual hand-painted models "AI slop"? Can this output even be called "slop" when it's fundamentally just the computer algorithm picking colours?

 

I'm not defending the posting of actual slop (there's plenty of utter dreck out there, man-made and synthetic, that whilst technically hobby-adjacent we could really do without) but I do think it's a bit more complicated than just "AI is bad" and dealing with it should be case-by-case rather than a blanket ban.

[HeadlessCross]

6 hours ago, Shinespider said:

I think it's a solid policy for the most part. This is a big forum with a diverse range of content, and I think a total blanket ban would probably be overreaching.

That said, I think the door should be open to localized bans. The Hall of Honour subforum *should* have a blanket ban of AI, for instance, imo. 

I don't care if the slop is diverse, I'm not consuming slop

[ThaneOfTas]

37 minutes ago, Evil Eye said:

 For example; just as an experiment I asked Google's Flow AI to rustle up some simple colour swatches for inspiring an Ork paint scheme. You can see the prompt I used and the output here;

I dont have any issue with people using it for stuff like that. I also don't think that people have any need to post that until they've actually started to work on it, at which point you've started your actual creative process. Hell, take those colours that have been generated and slap them into Impcat or something before you post it for feedback, that way you at least have done something that's worth commenting on and discussing.

 

If you want to share preliminary concepts of a planned kit bash, learn to photobash or use blu-tack if you want feedback. 

[Blissful Brushes]

AI is a tool that I don’t think most people understand fully and I would encourage people to do some research into how you could use AI to enhance your hobby experience. This is probably slightly tangential to the purpose of these guidelines, but what it is brilliant for, and what I use it for primarily is world building. I have a ~400k token background document that was partially built through Gemini that I use as a foundation for world building for my Luna Wolves and individual characters within my own little corner of the galaxy. 
It’s also pretty good at doing things like removing the background from images, or mocking up colour scheme concepts, so it has its place for the less technical of us.

I also made an app with it for my paint recipes which is proving handy.

 

In terms of videos yep deep six it.