Comment by jamiejones1

They're focused on making their models better at answering questions accurately. They still have a long way to go. Until they get to that magical terminal velocity of accuracy and efficiency, they will not have time to focus on security and safety. Security is, as always, an afterthought.