Yes, but Microsoft hasn’t put together that AI making mistakes, is perfect plausible deniability for intentional “mistakes.”