Valid questions, but neither deal strictly with autonomy. There are lots of unmanned airplane now, but many (most?) of them are not strictly autonomous - there is no explicit decision-making capability onboard. So:
1) We're already at the point where the machine can fly the airplane better than the pilot can without the machine. The machine is there to make it easier for the pilot to fly and the pilot is there to make the difficult decisions and deal with contingencies - at least some of them. In the past two years carrier launches and traps and probe-and-drogue refueling have been demonstrated without a pilot onboard. That covers two of the most difficult tasks (yes there are others) that are done in tactical airplane today. Automation is here and some level of mission-level autonomy is coming, whether the fleet pilot is ready or not. Right now it's expensive - the software testing required to flight-qualify a digital control system is formidable, whether automatic or autonomous.
2) These questions *are* being dealt with in the R&D community. There have been unmanned helos for a while now - look at the A160, Unmanned Little Bird, and Firescout, among others. Regardless of what one thinks about these programs, they demonstrate that the interface and C2 issues have been considered. While their missions are different from the H-60 you linked you can be sure that the human-machine interface and concept of operations is being considered - for fixed-wing, rotary-wing, no-wing, underwater, ground, etc. However, it's not a trivial problem and there isn't a simple solution that covers every platform and every mission.
Nauga,
who ran a little long