Preface: I have seen there has been a lot of discourse in different LSPDFR discords about vibe coding and the AI disclosure. I recently updated an older mod and had to fill out the AI disclosure. To be honest, I was a little confused. The wording was vague. There is no clear distinction between the 3 different tiers(minor, moderate, substantial). The suggestions below aim to improve that. Also some additional thoughts.
There are 4 main ways a developer can use AI to aid in their development of a mod: Ideation/Design work, Implementation(the actual methods/code), Debugging, Marketing(Description and Images).
Currently, the first 3 ways are represented with 1 option and the last option (Marketing) is represented with 2 options. This feels lobsided and where I think the disconnect is happening.
As an example: If the AI did a substantial amount of debugging which in turn changes a lot of code, I would think one would have to choose "substantial" for mod and mod content. Therefore, my mod would be slapped with the "Made with AI" tag. This gives users the impression in this scenario that I vibe coded this out with Claude Code and posted it on the forums with no thought behind it.
How I would improve the system?
My theoretical new disclosure system would look like this:
This gives each aspect of script development its own disclosure giving the user a clear picture of how the developer uses AI and whether the AI is an aid or the developer.
If this new system feels too invasive, I would understand that. However, I think the hesitation in telling the truth comes from what the consequence is if you did use a lot of AI; the dreaded black tag, "Made with AI". Many devs nowadays use AI as an aid within development. This system hurts those devs just as much as the devs who are using Claude Code to build plugins ("AI Slop").
If this system does not get implemented, I would maybe suggest changed how the tag is phrased.