One-Page RPG Jam AI Post

A look at generative AI tools in relation to the One-Page RPG Jam

Last time I said that the survey responses regarding generative AI deserved its own post, as the topic is ‘thorny’; i.e. one a lot of people have passionate opinions about. As if to underscore that, the lone comment on that post - a post explicitly not about generative AI - was about generative AI. So this time I’m going to attempt to tackle the issue, and how it affects the jam, in a rational and measured way (if such a thing is possible).

If you’ve spent any amount of time on TTRPG-adjacent social media, it seems as though everybody is firmly against the use of AI-generated images in any capacity, regardless of the provenance of the images or the technology behind it.

In my day job I work in a “creative industry”, and in pockets of the industry not all creatives are as against AI as the above described TTRPG discourse might be. That’s AI in the wider sense, not just generative AI. However generative AI is also a topic of interest, and I’ve seen people in creative roles look to generative AI as a tool to help in their workflow - not to wholesale ‘create’, but to generate assets or elements that they would otherwise spend time looking for or creating wholecloth, and this is especially appealing when those elements won’t be used in a final work. I’m being intentionally vague here, because I don’t use a pseudonym on the Internet.

I wanted to give jam participants the opportunity, through the survey, a chance to voice their honest opinions on the topic. To see if the prevailing discourse is representative of the community as a whole, or stems mostly from a vocal minority. I wanted to see this both to inform my decision on how the jam would tackle AI and because I was just generally interested in it, having seen different opinions of generative AI from different spheres.

---

Okay, to the survey. This portion of the survey was concerned just with generative AI - that’s text-to-image generation or prompt-based text generation. If you read the previous survey results post, you’ll remember that 232 people responded to the survey, 200 of which submitted one or more games to the jam. Those 200 represent roughly 40% of the total people who submitted an entry, but the following figures also include the 32 people who responded that they did not submit a game.

For images, 6% of respondents said they used generative AI to create images which they then edited, with several others saying they used it for generating some part of a graphical asset (a texture, an AI filter, etc.) and a single respondent saying they put an unedited image straight onto the page, and another used an AI-generated image from a stock art website, rather than generating themself. One thing to note is that the survey didn’t define ‘edited’ as that was left up to the individual to interpret; I imagine one would likely use a generous definition.

For text, we see similar numbers; 5% of people said they used AI to draft game text and 2% used generated text in the finished product. Additionally 8% of people said they used an AI chat bot to source ideas.

We need to acknowledge here that the options presented in this section of the survey allowed participants to select more than one answer - for example, if they used generative AI for both images and text - and also gave the option for a bespoke response. This means that some people opted to give a different response which in reality is close enough to one of the above responses, so the figures above are realistically one or two percent lower than the reality.

The advantage to giving a bespoke answer option is that people could comment when they had used generative AI tools outside of the narrow parameters of the question. Multiple people said they had used AI text tools to proofread and check spelling and grammar, especially if English wasn’t their first language. Several respondents said that while they hadn’t used AI-generated images in their submission, they had used AI for ancillary images, such as the game’s Itch.io page.

I am, though, spending a lot of my word count nickel-and-diming a few percent of answers. The big number is the respondents who said they did not use any generative AI tools in their jam submission - 78.9%. An overwhelming number of participants.

The next question in the survey presented participants with some statements, and asked them to select the most appropriate option - Never, Occasionally, or Frequently.

75% of people said they would Never “use AI-generated images in a game [they] create”, a number which climbs to 83% when asked about AI-generated text. Those numbers fell to 66% and 72% respectively when asked about buying games that used AI-generated assets. So it seems that people might be more open to others using generative AI, but wouldn’t use it themselves. However the numbers are still significantly against it.

Finally, I gave an open spot for people to write anything they wanted about generative AI. Unsurprisingly, there were a lot of comments. I read them all, twice in fact; once when I closed the survey and again when I came to write this post. They mostly exist on a bell curve from apathetic to militantly against, and the majority fell somewhere in a moderate middle-ground. There were, however, recurring themes. I won’t reproduce exact responses here, but here are the general vibes.

• A lot of people were fine with AI-generated assets in a free game, but against them being used in a commercial or paid-for product.

• Many responses were okay with text bots being used during the early phases of designing a game, but as one person put it - “I don't want to play a game someone else didn't even write“.

• A lot of people think it is against the spirit of the jam (or creativity in general). The word ‘antithetical’ was used more than once.

• Multiple responses said that any use of generative AI should be labelled or otherwise made obvious.

• Many, many, people were concerned with the ethics of it.

---

Regardless of where you fall on the bell curve, you cannot deny the ethical issues surrounding generative AI. The two main ones often cited by people who fall into the anti-AI camp are the environmental impact and the sourcing of the datasets the AI is trained on (i.e. theft). The latter is cited often in online discourse and was repeated many times in survey responses. Tools like Midjourney and DALL-E scraped images from the Internet to train their models, without compensating or crediting the millions of creatives they used. It goes without saying this is a big problem, but it seems to me this is a problem with the dataset method and not the technology itself. Adobe Firefly insists that it is “trained on a dataset of licensed content, such as Adobe Stock, and public domain content where copyright has expired” and, according to ZDNET.com, Adobe has compensated contributors to the dataset. If true, does this satisfy people’s concerns on this front?

The other oft-cited issue is the vast increase in electricity and water required to support the systems behind these tools, and the impact that has on the environment. We’re now way beyond my expertise here, and honestly beyond the scope of the jam, but on the subject of AI writ large, the environmental impact is something that certainly needs addressing.

---

All of the above is important background, but my main focus after all of this is how the One-Page RPG Jam should be handling AI-generated content. Not the industry as a whole, and not on a personal level, but what rules the Jam will have, if any, surrounding generative AI. So one can pontificate on the philosophy of AI, on the ethics of the technology and how it’s used, but ultimately there are practicalities that need to be considered.

For starters, how realistic is it to police a ban on AI-generated images or text? A lot of AI-generated images look obvious; they are uncanny or downright bad. Except, of course, the ones that don’t, and aren’t obvious. Would it be better to allow it, but ask participants to label it, so that others can make an informed decision? Prohibition, supposedly, doesn’t work and just drives the behaviours underground. Would it be better for the jam to have it visible so people can knowingly choose to not engage with it? Or would that just be a step to normalise it, another drop in the pond of acceptance in the wider TTRPG hobby?

I don’t, unfortunately, have the time to go through every entry to the jam and read them for fun, let alone checking for AI images or running the text through some sort of detection software, and asking people to report uses of generative AI risks a witch hunt and hurting people who didn’t use it but someone thinks they did.

But on the flip-side, throwing open the gates and allowing all uses of generative AI risks flooding the jam with low-effort and unappealing entries. Each year I’m amazed and immensely proud at the fantastic writing and graphic design that gets submitted. I don’t know if I would have the same response looking out on a sea of samey-looking games and images.

---

If you’ve read this far, I hope you can see that, even if there is something you disagree with or think I’ve missed or misunderstood, any decision I make for the jam isn’t some knee-jerk reaction to online discourse, nor because I’m deep in some tech-bro/luddite (delete as appropriate) echo chamber. With that out of the way, here is the rule regarding generative AI tools for the 2025 One-Page RPG Jam:

Generative AI tools may not be used to create any images, elements of images, or text used in a submission or a submission’s Itch.io project page, except for limited modifications to a translation from another language.

Firstly, that might need some unpacking. Secondly, I’ll explain my reasoning.

The above statement/rule feels a little legalesey. It means that you cannot generate images using an AI tool, such as Midjourney, DALL-E, or Photoshop's built-in generation tools, and stick them on your game. You cannot generate part of an image (like a character or object - then edit it or combine it with a non-AI-generated image. You cannot generate an asset, such as a texture or shape, and use that in an image. You cannot take a photo or other image and run it through an AI filter and put that in your submission. All of this applies equally to images you would make yourself from a prompt, and images generated by another - such as a stock image site.

Additionally, all of the restrictions above also apply to the Itch.io page for your game - the page, cover image, banner, etc. must all be AI-free.

For text, it's a little more complicated. You cannot ask an LLM or chat bot, such as Chat-GPT, to generate text for you from a prompt or re-write paragraphs of text you've written yourself. This includes fixing spelling or grammar errors; you can get one to help identify errors but not rewrite the text for you. You cannot ask a chat bot to translate an entire game or paragraph for you, but you can use it to double check a translation you've done yourself or another tool has done.

As with images, these restrictions on text also apply to text on the Itch.io page.

You can use an AI tool to help bounce ideas off or source information, as long as none of the words generated by the tool end up in your finished submission.

If you're using a tool or app that isn't a dedicated generative-AI-based tool but that uses generative AI within it, you can use that tool or app but not those generative AI functions.

---

I’m hoping this is a good middle ground. This isn’t an unenforceable blanket ban on any tool that has AI stapled to it as a marketing buzzword, nor is it opening the floodgates to low-effort slop. Ultimately, it’s largely where a lot of people already seem to be at. It would be impossible to keep everyone happy, and I’m sure there is a contingent who are disappointed that they can’t use AI to generate images, or even parts of them, for their game. To a degree, I get it. I’m not an artist, and I’m not hiring one to illustrate a free one-page game. I get why it would be convenient to type a prompt into a website and get an illustration, or even a background or something.

As many people commented in their survey response, the jam is about creativity. The one-page restriction has always been about making game design accessible in scope and the jam has always given people a chance to explore their creativity in a small, low-stakes environment. To enjoy working on something new, push their boundaries, and take themselves, at least a little, out of their comfort zone. If you don’t think you can draw, draw something anyway. If you don’t have experience writing, write something.

I outlined above an issue with disallowing generative-AI. As far as enforcing this rule goes, it’s the same as all the rules. If I see a jam submission that breaks a rule, I will remove it from the jam. Each year I remove a bunch of submissions, usually because they are not actually one-page TTRPGs but rather video game submissions or something. I will not be going through every submission with a fine toothed comb - the AI rule will be largely on the honour system - but if I see anything that obviously violates it, the submission is gone. As always, if anyone else sees a submission that violates the rules of the jam please report it, but if you’re reporting a game for breaking the AI rule please, please, please be very sure the game is utilising generative AI. I don’t want to have to wade through tonnes of reports against legitimate submissions!

---

This jam is as much about the community that has built up around it as is it the jam itself, and I welcome any feedback you wish to share before, during, or after the jam. Thank you to everyone who filled out the survey, both on the AI and non-AI front.

In an effort to make creating your own art easier, before the start of this year’s jam there will be a new #artists channel on the Discord server for people to discuss making art for their games, share their own artwork, and look for collaborators. I shall also be aiming to increase the number of links to art-centric resources in the jam’s resources section of the community forums.

Thanks all for reading, and I look forward to seeing all the amazing stuff you humans create during the jam!