I advise against using ChatGPT when dealing with factual matters. It can't actually tell whether what it's saying is true or not; all it knows which words are statistically likely to be next to which other words in the training data. It'll happily spit out complete nonsense and confidently present it as fact simply because it's a statistically plausible arrangement of words, and it's also somewhat biased towards telling the user what it thinks they want to hear in order to drive engagement.