MFS - safetypee has the answer, I think.
"Previous types did not have/use AOA, its only now that we rely on it that the failures are noticed."
I am convinced that the complexity of the current FBW systems leads to significant 'unexpected consequences' as designers add extra comparators and check logic to prevent malfunctions due to malfunctioning detectors, and each 'addition' adds its own failure complications such that driving to the root problem is extremely difficult. As has been said many times, having 'majority voting' was seen as the obvious and correct answer, but we appear now to be seeing 'majority errors' leading to isolation of the correctly functioning device with serious consequences.
Where then do we go? Do we add another sensor or 3 to the current group of 3 to give a 'casting vote'? Surely we unfortunately return to the need to have pilots who can
a) Easily wrest control from the malfunctioning system - DESIGN and
b) then fly the aeroplane as it should be flown until the wayward HAL can be sorted out - TRAINING.
The prevailing 'addiction' to the extremely clever FBW systems and an apparent undeniable belief in how 'wonderful' and 'infallible' it all is needs to change. We need to recognise that it is indeed 'wonderful' and 'infallible' when all works correctly, but when it doesn't, it can be a nightmare and push workload through the roof - and this does seem to be viewed as blasphemous by many.