We need to encourage the norm that AI is for personal use, and not for spamming others. If anyone wants to know what an LLM will output, we all know where to go to get it. It is immoral to pawn off LLM writing as your own. Doubly so, for trying to cover it up with intentional spelling errors or other "tricks".

AI is not so good at identifying a writing style by name.

Instead I feed it a few articles I wrote and tell it to follow the same style.

This gives way better results imho