When you create a new thread in the “Works In Progress” category, the “new post” template will invite you to disclose whether you’ve used any AI in your WIP.
AI Disclosure:
No — This project does not contain the output of Generative AI
Yes — This project contains the output of Generative AI
We recommend that every WIP disclose if their project contains content produced by generative AI tools such as LLMs, ChatGPT, Midjourney, Stable Diffusion, etc., even if you hand-edited it.
(For now, we don’t require that authors disclose their use of AI, but we might require it in the future.)
Choice of Games / Hosted Games / Heart’s Choice can’t publish art, code, or prose that was generated via AI, due to ongoing legal uncertainty around the copyright of AI-generated content. Your submission, including all artwork, must be created by a person or group of people who have the legal rights to publish it, and we have to provide them credit in the Credits.
It’s OK to use AI-generated images in in work-in-progress projects, as long as you replace the AI images with human-generated ones before submitting for publication. (If you do this, we strongly recommend that you disclose that your images have been generated by AI.)
Hmm. I approve of this but I am worried by the optics of it. Generally speaking, theres a (admittedly rightful) expectation that works that use A.I, particularly generative A.I is of a slop or low quality. As such, I imagine projects that do state their use will likely face a poorer first impression if not entirely glossed over because of that. And while that is to be expected when using A.I, particularly in large quantities. The FAQ also mentions using tools like Grammarly or other A.I for grammar purposes. I am in support of disclosing that. But I feel a binary, “This contains A.I work” and “No, it doesn’t contain A.I work” sort of lumps projects that make heavy use of A.I prose and those who have Grammarly in the background as one.
Wouldn’t it be better to have more options? I.E “This only contains Generative A.I art.” Or “This work was assisted by A.I grammar tools”
The template has invisible comments on it, like this:
AI Disclosure:
<!-- Please disclose if this project contains content produced by generative AI
tools such as LLMs, ChatGPT, Google Translate, Midjourney, Stable Diffusion,
etc., even if you hand-edited it.
Choice of Games / Hosted Games / Heart's Choice can’t publish art, code, or
prose that was generated via AI, due to ongoing legal uncertainty around
the copyright of AI-generated content. Your submission, including all
artwork, must be created by a person or group of people who have the legal
rights to publish it, and we have to provide them credit in the Credits.
It’s OK to use AI-generated images in in work-in-progress projects, as long as
you replace the AI images with human-generated ones before submitting for
publication. See https://forum.choiceofgames.com/faq#ai-disclosure for details.
To check a box, add a letter `x` to the box, like this: [x]
-->
[] No — This project does not contain the output of Generative AI
[] Yes — This project contains the output of Generative AI
<!-- If you checked "Yes," please provide additional detail below. -->
The key point is that last line:
If you checked “Yes,” please provide additional detail below."
Providing additional detail is definitely a good thing. We invite authors to provide detail/context on how you’ve used AI in your project.
I think if you write “this game was written entirely by AI based on a three-sentence prompt,” you’ll learn something from the reaction you get.
If you write “I used Google Translate to translate this from my native language,” that will get a different reaction. If you write, “I used Grammarly extensively to clean up my text,” that will get yet another different reaction.
I think this will be helpful for potential readers! I’ve noticed quite a few WIPs using AI generated images for temporary covers (at least I hope temporary) and have wondered if that was the extent of AI use in the project or if it was also in the writing.
Thanks a lot for contacting me and letting me know about it. My project doesn’t have any ai. I only wrote it using keyboard and a couple things that inspired me. My pic on the cover is an artwork I created in Pinterest so everything in the project I thought or sought about it.
I agree that disclosing the use of AI is important.
Does CoG allow the use of AI translator for words or fantasy words.
In my case, I found what Crown Prince in Arabic is online. But when it came to Crown Princess, I couldn’t find anything.
I knew that Crown Prince was “Wali al ahd” but I had to find a female equivalent.
I used Context Reverso and Google to find a translation that I think could work.
Which I settled on “waliyyat al-ahd”. For now, this is a placeholder until I update my WIP and get feedback from users.
And also what about fantasy word translations to find an equivalent Arabic name for a town or valley or fort, etc using AI for said translation.
So will a work that used a AI translator and later on gets reviewed by native speakers, will that be considered prose generated by AI which is not allowed.
I do sometimes chat to ChatGPT about my story. Mostly I use it to encourage me and butter me up so I feel inspired - which it is very good at, I am slathered in butter, a veritable croissant when done. And I asked it for era appropriate perfumes and the names of Regency hairstyles. It also translated a letter to French for me. Is it necessary to disclose this? Is it allowed?
Because I do think we should perhaps define rules here. I get and agree that AI shouldn’t be writing for us and that we shouldn’t use AI art. But is it wrong to talk to a Chatbot? Is it wrong to ask it for references? I’d like some clearer definition if that’s okay.
The OP says that it’s not a requirement to disclose, but a recommendation. So it’s not “necessary” for a WIP thread, but games can’t be published if they contain LLM-generated prose.
(I’d personally recommend using more reliable means of research for historical details.)
If you use it to feel well about yourself to feel inspired, think about how much disdain you might earn when people learn you use genai for writing. That’s a writer’s block in the making.
I’d say skip the bot and see about finding a nice community of actual people gestures around the forum most people here are nice.
I think the rules have been pretty clear. There are some more details here about translators.
I’m sorry if my comment about research was alarming; I meant it as a personal recommendation not a rules-statement, because genAI is pretty low accuracy.
There’s a thread about general AI discussion here:
Thank you. Judging by the reaction here, I should avoid AI completely, even outside writing because it gets nebulous. I just talked to it a few times, I didn’t realize this would be so vehemently frowned upon.
I’ll use Google Translate as a translator instead (for that one letter) and remove the hairstyles and scents just to avoid any issues.
I’m not sure if using AI for research is frowned upon to the point of needing to remove this, but I’d recommend at least re-checking whatever results it gives you manually. Not everything it tells you is automatically 100% correct.
Oh, I’m aware, but thank you. I thought it was useful to get a list and then check from there. Writing historical fiction does make research a bit difficult sometimes, no book that I’ve read or research I’ve done compiles perfumes of that era. Or hairstyles! ChatGPT can show you where it got those references.
But I won’t use it again. I think the subject is clearly way too contentious and I don’t want to fall afoul of COG or its readers.
You might know this, but since you mentioned wanting to avoid AI completely the guidelines classify Google Translate as a form of AI usage.
“Generative AI” refers to artificial intelligence systems that create new content (text, images, music) by learning from large datasets. This includes large language models like ChatGPT, Gemini, and Google Translate, and image generation models like DALL-E, Midjourney, and Stable Diffusion that create new outputs based on training data.
If you use an AI grammar tool like Grammarly, please disclose that, too.
If you can find a non “AI translator”, that might be a better bet.
