This seems like it doesn't really answer OP's question, which is specifically about the practical uses or misuses of LLMs, not about whether the "I" in "AI" is really "intelligent" or not.