Rule Specification about non-AI Images/videos

Since Civit is an AI site, it should be expected that it would feature AI generated images. However, there are use cases where displaying non-AI content is done to show the functions of AI technology, such as img2img content, video2video content, and what I planned to post, AI_img_2_3D_print content.

When I was asking a moderator over Discord if I was allowed to post an irl photo of my 3D print to the page of the lora that created the OG image (which had been converted from an image to a model using hyperhuman.deemos.com), the mod said that it was not allowed. When I asked further if it was allowed as an article’s display image, I was again told it could be removed. When I asked further why img2img and vid2vid content and articles could show irl content as a comparison, the mod could not seem to find a rule that distinguishes acceptable from inacceptable non-AI posts. The only thing they found was the Civit Safety Center quote “Only post original images that you generated yourself“, a rule meant to dissuade content duplication. Technically, no self-taken non-AI shot is even banned by this logic, because this rule does not specify anything must be AI generated, only that the user must generate it, yet the moderators treat it as contraband anyways.

What my story is trying to illustrate is that the rule against non-AI content is purely unspoken or at the very least hard to find. It is further moderated using acceptable use case exceptions that also go unspoken. Can we please have clarifications on the rules about:

• What non-ai images/videos are acceptable as posts?

  • Would it need to be posted with the relevant AI content, or can you just explain it’s relevant in a comment/post text?

• What non-ai images/videos are acceptable as posts to lora/checkpoint/other resource pages?

• What non-ai images/videos are acceptable as article content?

• Are there any special considerations for content based on ai material (like my 3D print post, paintings inspired by ai art, etc.) as opposed to ai content based on non-ai material (img2img, etc.)?

• Any other considerations?

Also, if there is any existing rule about this, is there anything here that has not already been addressed?