Understanding incidents and accidents can be biased by the initial framing.
Statements using ‘crews’, ‘pilot’, ‘they’, are more likely to be interpreted as ‘pilot bashing’, human factors at the sharp end, opposed to considering alternative contributing factors
From the examples above:-
Weak auto thrust design, the system could go to sleep; the designer was surprised.
Weak FD switching logic, mandated use of FD for takeoff; the operator was surprised.
Multiple probe icing beyond the conditions assumed by regulation; the regulator was surprised
In these events and others, the expectation (after the fact) is that the human should have managed the system deficiencies. Our, operational expectations are influenced by hindsight, whereas design and regulation have to anticipate.
The expectation is that humans will manage malfunctions in those systems which were fitted to ease workload, minimising ambiguity, to provide a higher level of safety than the human alone was capable of.
Catch 22: After recognising human limits, additional safeguards were provided; however these required the human to manage failures in the same systems designed to overcome human limitations.
The industry has a high level of safety, such that accidents have few if any common factors, except the human; thus inappropriate focus on the human. It is time for safety management to appreciate uncertainty in operations and its effect on human performance; don't start with the human, consider the overall system.