Proving an AI can create information borders ridiculously close to the fundamental halting problem. Any information produced by AI needs to be parsed and curated both in input and output, by a human who can determine if the information provided is feasible. Trusting an AI provides any real answer is wrote incompetence with computation.