I'm curious about how taking the human being out of the equation would actually be accomplished.
Boiled down to some simplistic ideas-
Take the pilots away.
Theres a computer. With a program. Will the program be written by a computer, or human?
The aircraft will be refuelled. Will the refuelling be done by computer, or human?
The aircraft will be built and maintained...
I could go on, but the general idea is there.
The computer should only augment the situation, not replace. As has already been mentioned, Humans have notably been the saving factor (yes, also the weak link at times) in a number of very prominent events.
Is it the reality that somewhere along the line, it will be tried? Yes. But insidiously, rather like our gradual decline in standards. UAVs doing the reconnaissance, crop spraying, banner towing, combat, to name a few.
There would be a level of success sufficient to make it commercially viable, with acceptable levels of risk and loss built in. What level of redundancy will be needed to ensure that multiple system failures won't automatically result in losses?
To go along with wholly automated flight, there would have to be a dramatic change in systems capable of last ditch efforts at saving the SLF from a grim end.
I don't trust computers 100% (I'm an ex engineer), but neither do I trust humans 100% (ex engineer and current pilot), nor do I trust myself 100% (you should see me on day 1 after a break! I won't even mention days 2 thru 6)
I wonder how the Apollo landing, Sioux City and the Sully Ditching would have turned out with total automation.