I have a twofold concern regarding automation. First, I am concerned that pilots will be too dependent on automation and not maintain proficiency to perform when needed, as demonstrated by Korean Asiana crash going into San Francisco and, second, the sophisticated auto fly system may not go through the same rigorous validation as the standard and proven fly-by-wire systems. The Boeing 737 Max crashes, and probably some Airbus incidences, demonstrate two ways automation can go wrong. First, apparently two angle of attack sensors failed (on two different aircraft???), and second, even when recognized, overriding the auto system was complex and was not implemented in a timely manner, Pilot training may also have been an issue. Where was the system redundancy? The B-2 has a quad-redundant, identical path, flight control system. It has four independent sensors for angle of attack. These are compared for reasonableness (all four must fall in an acceptable range) and voted on (e.g., the high and low ones are discarded, and the two remaining are averaged and used). A duel system, ala 737 is never acceptable. If a sensor is starting to vary and is undetected by built in test, the flight control system cannot determine which sensor is correct.
This is a subject that I am quite familiar with and have considerable knowledge about. My background knowledge comes from flying the C-141 in the early 70s, and afterwards being the Avionics Controls and Displays manager for the B-2 bomber from proposal to AF operational. At that time the C-141 was quite advanced with full auto land including auto approach and auto throttles, CAT 2 capability, 100 ft ceiling and ¼ mile visibility. Other than that it was basically steam gage instruments but with steering commands. I NEVER flew an automated approach and landing. I always did it by hand. I did note that while checking out the instrument landing system couple that it flew more precisely, with the steering commands always centered, than I could.
In my role on the B-2 I was engineering manager responsible for design requirements, development, test, and delivery of the hardware and software of the controls and displays. As far as I know, the B-2 was the first fully integrated cockpit with multifunctional displays, certainly in the US and maybe in the world. I don't think the F-18 had multifunctional display and certainly did not include the engine instruments or threat indicators. I do not believe the 757/767 had multifunctional displays but rather were replacements for the vertical situation indicator, horizontal situation indicator, and engine instruments.
While the B-2 had, for the time, sophisticated processing capability, four identical flight control computers in a fully fly by wire system, and two identical flight management processors. I did have a couple of advantages in implementation over commercial applications. The AF is run by pilots, who, at the time, were not interested in automated flying outside basic autopilot functions. In fact I offered to provide fully automatic aerial refueling station keeping, a high work load environment. They wanted nothing to do with that. In addition, commercial operations typically are in the business of moving passengers/cargo from point A to B with training squeezed in somewhere. The military war fighters business is war and since there are few wars going on, every flight is a training flight so they become very familiar with normal and abnormal operations of the equipment.
Some interesting aspects of the B-2 display system:
Four identical full color displays are provide to the pilot and the mission commander (right seat)
Formats are all completely interchangeable within the four and within the two stations. Only limitation is the Vertical Situation Display must always be displayed on one display.
All avionics functions are integrated into the controls and displays subsystem. There are no dedicated display or control heads except that mandated by mission requires, e.g. standby flight instruments, and engine instruments (primary instruments were on the displays) which are required for base escape when avionics is not full up, etc.
All avionics control/display functions are rated by criticality and frequency of use, for example Identification Friend or Foe (IFF) emergency is critical and Mode 3 is frequently used, so both are top level, whereas other IFF modes are less critical or frequent and may be at a lower page.
I had a letter on the B-2 cockpit that was published in Aviation Week and Space Technology. Since we were a Black Program, that letter not only had to be approved by my management to the VP level, then approved by the AF Program office, but also by a Congressional subcommittee.
Attached is view of the B-2 cockpit, the outline area are the instruments I was responsible for.