DL makes valid points about the automation / human interface. However, I'm not sure that this accident illustrates his point all that well.
As far as this passenger can tell, airlines prepare for dropout of automation by providing lists of actions to be carried out; in this case, the UAS checklist. These action lists allow the pilots to keep the aircraft safe for a short period while their physiology adapts to the new circumstances.
It deeply shocks me to read the CVR transcript and NOT see the AF447 pilots going automatically into that sequence. Surely there must be something in the displays / warnings that triggers the pilots into recognising why the A/P has dropped out? Concentrating on the UAS actions might have given them time for hands to be steadied, or even for the ice to clear and the speeds to come back to normal. Neither of them said anything along the lines of 'Help, the displays have all gone black' at the start of the trouble, so I assume the information was available to them.
The plane wasn't in immediate danger, apart from the fact that it was ploughing straight through some very bad weather. It was doing this because that's what the human crew had elected to do, not because the automatics had misread the situation. This was not the most low-risk strategy available to the Captain, but he had not only chosen it, he had also reduced the chances of managing the effects of the weather by choosing to absent himself at the critical time. Anxiety about the plane's course seems to permeate the younger F/O's words, and perhaps contributed to his later reactions.
The human /automation interface is undoubtedly a factor in this accident, but I can't see that it's the principle one. The finger has got to point at Air France and their training /management of aircrew, and quite possibly at their culture of performance measurement and rewards. Or is this standard of cockpit discipline the norm in other airlines, and I'm just an innocent to believe otherwise?