Yes one of the examples was when you are not entirely sure of how to carry out some task and ask AI a method, it will happily give you something which looks a valid method but has lots of errors or even just one which actually makes it useless and risky.
Quite creative at looking and sounding good but not completely correct, which is useless and dangerous if you publish AI's offering without checking.
Quite creative at looking and sounding good but not completely correct, which is useless and dangerous if you publish AI's offering without checking.
It's not the least charm of a theory that it is refutable. The hundred-times-refuted theory of "free will" owes its persistence to this charm alone; some one is always appearing who feels himself strong enough to refute it - Friedrich Nietzsche