Why do you trust something that is LLM-generated is authoritative?

Even if, on the most good-faith scenario, someone did all the research them told an LLM to add more pizzazz for some reason, there is still a chance that the LLM hallucinate a lie here and there that the author might have skipped when proof-reading if they proof-read it at all.

I can't trust a machine specialized in generating convincing bullshit.