large language models (ai) cannot be used to "verify" information. it's just a tool. i think people didn't understand that either.