The Cargo, Not the Ship

As Marshall McLuhan said "The medium is the message.". You cannot separate the cargo from the ship in this case because the functioning of the ship/AI unavoidably contaminates the cargo/message.

So, what does the medium of LLMs tell us about the message passed through it? What are the sub-textual aspects of messages with obvious AI tells? None of them are positive: risk of factual errors from hallucination, risk of it being a low effort message, risk of the message being subtly and unintentionally biased by the biases of the AI, the list goes on.

Just change the analogy from cargo/ship to item/package. If the packaging is damaged I wouldn't expect the item inside to be in good condition.