Yesterday's report on the causes of the crash of Air France Flight 447 is incomplete, and the reliability of the investigation's findings will never be without question. But the broad conclusion is probably all-too true: the computer had something to do with it.
Well, it's more than likely that many automated systems played roles in the accident, but it seems to have started with the plane's speed sensors yielding incoherent readings. The bad data may have tricked other systems in doing things that weren't in the best interests of staying aloft. Analysis of the wreckage suggests that the pilots tried to land the plane on the water, as if they'd eventually wrested control from the computers, only too late.
Automation is a wonderful and frightening aspect of our lives, isn't it?
From toasters to power steering, we rely on machines to do things that our ancestors either had to do manually, or couldn't do at all. Computers have made that automation much more pervasive: we expect our hard drives to preserve our family pictures, just as engineers at nuclear power reactors depend on software programs to make sure the fuel rods don't overheat.
I'm a firm believer in habit and routines as powerful brand attributes, and computerized automation is a key strategy for delivering it. Subscriptions, 1-click purchasing, product suggestions based on past interests, and automatic upgrades/renewals are as important as any conscious attachment to a product or service, perhaps even more so. From a functional perspective, "smart" devices can outperform "dumb" ones, generally, which means they can charge more for delivering better experiences.
Only now I'm becoming more aware of the downside to relying on something unconsciously, and it's called dependence. An automated function requires an abrogation of authority, or power, from beneficiary to provider. Because the platform on which I type this essay has all of the editing functions represented as buttons, I've never had to learn how to program in HTML. I rely on cruise control in my car to keep my speed constant, instead of learning to to do with foot and pedal. Forget to send me the subscription renewal, and I might neglect to prompt it myself.
When the pitot probes fail, the plane is sent into a literal tailspin.
We're not always aware how much we give up more in order to get more from our automated devices and services; making automation smarter, or more pervasive, doesn't necessarily mean it's any more reliable. Greater responsibilities increase the range of possible failures. Even if machines achieved consciousness, they'd still be prone to making the same mistakes in judgment that we make, wouldn't they, or it, or whatever? God forbid machines developed attitude, like HAL9000 did in "2001: A Space Odyssey."
This condition in our lives is what fuels some of the interest in the Maker movement; their premise is that we'll find greater fulfillment from our lives by taking active control of them. That means making stuff instead of outsourcing it to someone or something else.
I wonder if events like the Air France crash (or any of the lesser, daily instances of automated things failing to perform) will get people thinking about the trade-offs they make in their lives between control and benefit?
- Could a next high-tech gizmo purposefully put more user controls back into users' hands?
- Maybe a subscription service could solicit more engagement, more often?
- Imagine automatic monitors, sensors, or services that required active, conscious involvement from operators, not less?
At least when it comes to critical stuff, like managing ICBMs or flying airplanes, maybe we don't want HAL to be able to do it?


