Proof. I am asking for anything for you to back up what you are claiming.

All of those things require context. Something LLMs cannot ever understand; it is a hard limit of the statistical analysis that LLMs use.

Edit: also, here is proof of cognitive decline brought about as a result of feeding LLM outputs back into LLM models: https://bmjgroup.com/almost-all-leading-ai-chatbots-show-signs-of-cognitive-decline/