On the topic of ML/LLM (so-called “AI”), let it be known that I consider all outputs of one whose input (“training data”) contains works by me derivative works of those which therefore require full attribution and licence compliance. ML/LLM are merely lossy compression combined with something like a compiler to recombine and re‑extract (see and here). Anything else is a licence violation (which, when done commercially, is a felony).

In particular, these interpretation guidelines shall be followed for any work I publish under Creative Commons licences.