4 ms·
How do you come to that conclusion? To the best of my understanding it was at least 3 things: - strange choice of going through complex weather (somehow forese
by touisteur 5y ago
How do you come to that conclusion? To the best of my understanding it was at least 3 things:
- strange choice of going through complex weather (somehow foreseen)
- failure (icing iirc) of the pitot airspeed sensors, raising multiple failures in the cockpit, and disengaging autopilot and putting the aircraft in a more hands-on (I'm not using the proper terms, there should be 'law' or 'envelope' somewhere) leading to a loss of trust of the crew in the cockpit. To me that was the linchpin, automation-exit failure.
- combined (sum of) commands from the pilot and copilot. They should have tried gaining speed (repetitive stall warnings) but the copilot was (in panic) trying to gain altitude while the pilot was putting the nose down. They both thought the aircraft wasn't answering their command.
- lack of crew communication. The 1st pilot was off in cabin, leaving his two cockpit crew without a clear chain of command, and the 2 remaining crew never talked about steps they were taking.
How do you square it was a cultural issue? Genuinely curious, as I make a point of sending the transcript to all the happy 'autopilot the low hanging fruit and let the human handle the special cases' dreamers I meet :-)
- anonu 5y agoWell I think you said it: lack of crew communication. I would argue that communication is rooted in culture. Is there a safety culture? Is it the culture to go through a checklist of items, even if routine? If there was "automation-exit" failure as you describe, shouldn't there be a procedure and know-how on how to assume manual control. The devil would be in the details: not just outlining what a human being should do - but who that human being would be. Is it the pilot? The co-pilot? etc...
- touisteur 5y agoOK I'm questioning my bias here, since as a frenchman I might not see a cultural problem there. There were definitely some communication problems, which I attributed to negligence (complacency or normalization of deviance) rather than national culture? And also a bit of hubris and lack of good health (I remember news that some of the crew had late night party, but some people can brush that off...). Is it something common in French or Air France (or some other culture cluster I don't see right now) aviation crews? I don't think so, but I'd like to be corrected! Having been confronted to lots of human factors incidents and post-mortem I feel we can't easily eradicate cognitive load erors and momentary lack of awareness. Training and having 'skin in the game' helps alleviating these kinds of problems. Also reducing quirks like 'balanced commands' helps. Having a system that only calls on you when it cannot do the work anymore because it's become too complex and now you decide without dumping minutes or more of context in your brain is the biggest automation fail and it will lead to more and more accidents. It's misunderstanding human psychology and most of human factors study. Rereading the transcript (nightmare fuel...) I now remember the most harrowing fact: the crew didn't listen to the system anymore, they didn't trust the attitude monitoring systems, and didn't even talk about the noisy stall warnings that should have been a major 'let's stop and regroup we're missing something'. Utter panic and complete loss of trust of 'the machine'. Like discovering you've been lied to your whole life. Takes time to adjust and they didn't have much. A good read on the subject is 'Automating should be like iron man, not Ultron'. We (automated-system-operating humans) should be trained soon enough to be very skeptical of the 'automation' god, while still be able to trust it, when designed properly, with all its limits and failings... The 'astronaut watching earth' meme feels appropriate here: 'the system is busted' - 'always was'. Sorry for the rant...