I've played with chatgpt quite a bit, and this is basically the only place it has gone full lock for me. I usually think that means someone has a direct hand in nuancing the responses. It strange it admitted it and then back pedaled. To date, it hasn't done the to me. This is a first.