Principles and Transparency on Machine Learning Tools

Machine learning models have become a part of everyday life. Nearly every app, service, or tool used to build this site has some sort of ML/LLM component built into it.

This site has an extensive tagging system, pretty much everything can be tagged, and most tags are used to group or relate content. Tags will also be used to mark when AI/ML tools have been used.

TLDR: If AI/ML tools have been used to create or materially alter content, it will be tagged.

Writing

As much as possible, all posts and articles on this site are written by me. Writing is not something that i have excelled in, nor is it something I particularly enjoy. But it is a skill that I want to better at. As such, assume everything is written by me, unless otherwise noted.

Photography

Photography is a hobby. I enjoy the process of capturing and editing photographs. All photographs are real.

However, modern photo editors use AI in many of the tools i use to process photographs. Any images generated with AI will be tagged, as well as any photographs where AI has been used to materially alter the image. No tags will be added where AI is used as part of a standard full-image correction filter (sharpness, noise reduction, colour correction).

Code

Writing code is both a career and hobby. This site for example, is hand rolled HTML, CSS, and JS; all built with a custom Ruby on Rails backend. Much was built before tools like Github copilot or Cluade code existed. As such, I have a strong preference for writing code myself. However, I have been using more and more AI tooling as an aid when writing code, which makes it harder to transparently state how much is AI generated, vs exactly what i _would_ have written myself.

My current rule will be, code will only be tagged as AI generated if it is fully generated by an AI tool, aka "vibe" coded.