Exactly what you want during a disaster - a chatbot confidently hallucinating nonsense answers to your questions when you can't verify them.
I doubt this is what they've done here, but not all AI chatbots have to be thin wrappers around a large language model.
And I'm the hysterical one?