How can you tell if the AI isn't hallucinating in these matters? It might be okay for summarizing texts, but for everything else, not so much.