Presenting raw LLM output as your own writing in a human-to-human context is rude and antisocial, not to mention incredibly transparent.