ChatGPT does not know what medical misinformation *is* and is therefore fundamentally incapable of being programmed against it. They could program it to avoid language that is contextually associated with medicine but there's way too much overlap with a variety of other subjects.