I'm not going to get diverted into the causes of the crash you mention, though I believe there are logical explanations behind it with the control logic that existed at the time.
The limitations imposed by Airbus's flight control system in Normal Law are more than adequate for all manouevres that may be required in operating the aircraft. They can be overridden if absolutely necessary, though it's not a particularly simple thing to do.
There is nothing sinister about the actions of these flight control computers; they do not fly the aeroplane for you (autopilots have been doing that since WW2). They simply have used rather basic computer technology to prevent sloppy handling from allowing the aeroplane to stray outside the manouevre envelope. Those limits are not significantly different from those of any commercial aircraft.
The computers make no 'decisions' as such. They simply monitor various parameters and, if any one parameter is approaching a limit, control inputs are automatically applied to attempt to keep that parameter within its designated limit. It is a simple 'if parameter X=this, then control input Y = 3 bananas (or whatever). This is a world away from free decison making in a constantly changing multi-dimensional environment.
And I repeat the point I made earlier: no commercial aircraft with pilotless flight decks are in development or currently proposed. All the aircraft that are envisaged at the moment will continue to be manned by two pilots. Given the life-cycle (now and envisaged) of modern aircraft, that suggests that piloted commercial airliners will remain the predominant species in service for at least the next 50 years.