Who is the arbiter of the fact checking algorithm? If we use AI can we really trust the outputs are hallucinated?