I agree that it's very useful, but I'd be careful about "fact checker". GPT is perfectly happy to confirm falsehoods as facts and hallucinate convincing lies. A good fact checker verifies from multiple sources and uses critical thinking, neither of which ChatGPT can do.

Plus the fact that no AI of this kind can explain why it produced the output it did, it denies any amount of accountability/dependability. Maybe it's my academic/scientific background that shaped me this way, but using an information that can't be sourced/repeated as experiment/proven from axioms and theorems is a liability and often a cause for gross negligence or misconduct.

It strikes le that most people don't see that as fundamentally problematic.

Very true, however, you verify "facts" by doing research and crunching data, you do that by Googling and searching, and this AI is just another way of searching and presenting that information.

Caution is always required.