your referring to protocol, not something thats technically challenging and could be done today between ATC and an on board computer.
moving to related but slightly different territory, the tech press last week spoke of the automated Google test cars having accidents. apparently they've had a dozen over a million or so miles, at first reported (by the lay press mostly) as a bit crap. turns out none of them were caused by the automated car, they were all the result of human error. the problem with automation is that is doesn't handle exceptions and the unusual very well. the positive it deals with routine, mundane really, really well. so what you have is a situation where automation is held back because the wetware around it cant be trusted, though we oddly don't have faith in the automation rather than ourselves. in an environment with a lot of external variables, automation is going to have difficulty. keep it in an environment with few variables, it will run very smoothly (cf DLR)
to bring us back to air travel, landing on the Hudson is not something a automated aircraft would "think of". however, in principle, had it a coded routine to deal with a situation with an emergency landing, and say directed to a flat piece of land/river to land on, it would conduct that landing very well - the process changes little. meanwhile, short of having HAL in the cockpit, an automated airplane wouldn't decide to dive into the alps. one day, probably not too far away, when we've accepted automated cars, we'll have all transport automated. the rail could be already if it wasn't for unions.
I've not got a clue what you're going on about.