Golden Ritual #1: My First Experiments with AI

From trying to generate Japanese rope bondage to exploring black leather, polished gold, and the limits of AI.

My first experiments with AI image generation started with something familiar: Japanese rope bondage. I wanted to see whether I could create Kinbaku and shibari images using the commercially hosted tools available to me.

That turned out to be considerably more challenging than expected. Rope, restraints, and bondage-related prompts quickly ran into the models’ content restrictions.

After repeated attempts, I changed direction and returned to one of my Beyond Rope aesthetic: glossy black leather and wide, polished gold restraints. That opened up another possibility, although getting the image right brought its own problems.

Golden Ritual #1 grew out of that process: a generated portrait, refined through several attempts, then brought into motion with AI-generated sound.

AI-generated Japanese woman in a black bodysuit with wide gold neck, wrist and ankle jewelry, kneeling in a Japanese interior.

Golden Ritual #1
AI-generated image · October 2026
Select the image to open the full-size lightbox.

Trying to generate Kinbaku and shibari

The original intention was straightforward: try generating Japanese rope bondage imagery. In practice, the commercially hosted models I used often treated Kinbaku, shibari, and related fetish imagery as sensitive subjects. Prompts could be refused before I had an image to work with.

It wasn’t only explicit imagery that caused difficulties. Even a fully clothed adult wearing polished metal jewelry could trigger a refusal. Changing the wording and trying again became a substantial part of the process, without reliably getting me closer to the rope images I had set out to create.

Rather than keep spending my time on those restrictions, I decided to take the experiment in another direction.

Turning to Beyond Rope

I returned to a combination I’ve already explored in real photoshoots through Beyond Rope: glossy black leather and wide, polished gold restraints. It gave me a familiar starting point and room to develop a fantasy image without making Japanese rope bondage the central subject.

The change of direction helped, but it didn’t make the process effortless. Content restrictions could still interrupt an attempt, and getting an image generated didn’t mean that the details matched what I had in mind.

Getting the gold right

The gold restraints were a particular challenge. Their shape, placement, and fit needed to look deliberate. Getting those details right took repeated generations, creative prompting, and further image refinement.

It also took a rather unusual solution to get the gold jewelry to fit exactly as I imagined. There was quite a distance between asking for polished gold restraints and arriving at an image I wanted to keep.

That process became Golden Ritual #1. Once the still image was ready, I used it as the starting point for a short AI-generated video, with AI-generated sound completing the experiment.

Golden Ritual #1 in motion

Golden Ritual #1 | Oct. 2026 · AI-generated video and sound

 

 

 

Production notes

This project combines AI-generated imagery, motion, sound, and text, with the concept and creative direction coming from my RopeMarks work.

  • Original image: FLUX.3 Image.
  • Image refinement: ChatGPT’s image tools. The exact model version was not recorded.
  • Video: Seedance 2.5, using the generated image as the starting point.
  • Sound: AI-generated with ChatGPT Work.
  • Text: Generated and refined with assistance from ChatGPT.

 

The Real Thing

The polished gold in this experiment draws on my original photoshoots. This selection will show the work that inspired the generated image.

Where this leaves me

I started out trying to create Kinbaku and shibari images. The limitations of the hosted models led me back to my Beyond Rope aesthetic, where a different idea began to take shape.

Working through the refusals, repeated attempts, and problems with the gold’s fit became part of making this piece. The result is a first experiment in combining generated imagery, motion, and sound, with a visual starting point drawn from my own work.

I’m curious to see where the next attempt leads.

 

  RopeMarks | 2026