Are machines more reliable than humans? In the wake of the tragic Germanwings crash in the Alps which took 150 lives, suspicions that it was an extended suicide by one of the pilots have brought about suggestions that human pilots should be replaced by automats and remote control. It is technologically possible, at least in theory. A big part of most flights are already auto-piloted and human error causes a large percentage of incidents. But are pilots really only about taking the plane and passengers to the destination?
Malcolm Gladwell spends a chapter in Outliers (Little, Brown & C:o, 2008) on plane crashes, and he gives some examples of how the human factor as the cause (but he doesn’t suggest pilots be replaced by machines, rather he talks about how communication between the people involved can improve). Gladwell also recounts an event where Sri Lankan flight captain Suret Ratwatte had a passenger who took seriously ill on a flight from Dubai to New York. It was an Indian lady from a small village, her husband spoke neither English nor Hindi, only Punjabi. It was probably a stroke and the lady needed immediate hospital treatment. Ratwatte found himself over Moscow but had little trust that the poor Punjabi couple would be treated appropriately by the Russian authorities, so he decides to land in Helsinki. Because this happens early in the flight, Ratwatte’s plane is heavy with fuel, but dumping it would demand a detour that could further jeopardise the lady’s life. So he has to take down very carefully with a too heavy plane in an airport he has never been before, at the same time communicating with doctors on the ground and among his passengers, the Helsinki ground control, his flight staff, the other passengers and his superiors on the ground. In the end, the landing went well.
Perhaps a machine will one day be able to do something like that: make complicated decisions, communicate with several people at the same time (handling both emotions and facts), at the same time pushing complicated machinery beyond the limits of what it was built for. And perhaps some of those difficult decisions could also be made by people on the ground, remote controlling the machines that fly the plane. But even if all that becomes possible, were you a passenger on that flight – would you rather have the person ultimately responsible up there in the air with you? Or somewhere else, far away from the situation?
Humans have one unique feature over machines. A killer app: Responsibility. We humans can be held accountable for our decisions and their consequences. Machines can’t. That makes a world of difference.
(For the reader that appreciates irony, I write this blogpost on a plane)
