Isn't this the truth; Air France 447 is a classic example. The pilots continued to attempt to climb with the aircraft in a full stall.
With Steve's indulgence....... Roddenberry wrote about this in the 60's...... people using advanced technology, but retaining enough detailed knowledge to fix/patch/modify those same systems at will. Checking to see if I was spelling his name correctly..... turns out he flew 89 combat missions in WWII. He later was writing sci-fi for television, but chances are he knew a thing or 2 about fix/patch/modify and flying by the seat of his pants, albeit in another era.
Air or sea, today, I'd argue that it is probably more critical now than ever to have people that are intimately knowledgeable, and fully engaged with, the details of these complex systems. Otherwise, you're just a passenger.
While my preference is usually to see a competent human in control, at the same time, I have trouble understanding why these systems (esp. non-military) don't have certain KEEP OUT defaults programmed. We had a BC ferry a few years ago that hammered into an island in the middle of the night (I recall it as something like 50 miles long - definitely not small). Ship sank, with at least 2 lives lost. That ship is still at the bottom of that sound, last I heard.
Multiple ships moving in close quarters is one thing, complexity wise. OTOH, what reason could there be for heading at high speed on a collision course with a large land mass ? Land masses normally don't move, at least in human time-scales.... With today's systems, collisions like the ferry one I just described, and possibly the Concordia, might be avoided by having at least several over-ride codes required. Modern telemetry can easily report safety-overrides to centralized control, allowing another level of oversight, and the possibility of dead-man control.
Rgds, D.