This chapter is from the book
This chapter is from the book
This chapter is from the book
Can Technology Change the Odds?
An important question is whether we can use technology to automatically prevent accidents in complex systems, and if so, are these measures a net positive force? Technology is being used for operations in more businesses every day. Common examples include automation of production processes, automation of customer service functions, call center support systems to help make operators "smarter" and more effective salespeople, and information systems that monitor key variables constantly and warn managers when limits are exceeded.
We want to believe that if we program operations and response, we can ensure standard quality and minimize or prevent mistakes. The reality is that business, broadly speaking, is not as far along in this regard as those businesses that must use technology to operate at all.
For example, Airbus Industries pioneered "fly-by-wire" and first introduced it in commercial passenger aircraft in the A320. Historically, the pilot's yoke (or stick, depending on the aircraft) was physically connected, via cables, to the ailerons and elevator, the primary control surfaces for roll and pitch of the aircraft. As airplanes grew larger and heavier, hydraulic actuator systems (something like power steering in automobiles) were connected to the cables to make it easier for the pilot to control the airplane. Even with a hydraulic system, pilots physically feel a direct relationship between the movement of their hands and the response of the aircraft.
Fly-by-wire removes the physical connection, with a joystick generating an electronic signal that is sent to actuators that drive the movement of the control surfaces. Fly-by-wire makes controlling a large passenger aircraft akin to playing a video game, literally using a joystick to control the aircraft attitude and direction. For pilots, this was a major technological leap that was not necessarily welcome since the "feel" of the aircraft is artificially induced in the stick by electronics and the response of the airplane may be limited by algorithms and parameters set in software.
For aircraft engineers, this technology simplified construction and maintenance and potentially enhanced safety. It meant that a computer could be put in the loop to limit what the pilot can command the airplane to do. This is an attractive capability providing engineers the ability to actually limit what the airplane will do rather than just writing a manual that warns operators not to exceed certain parameters. To improve safety, Airbus aircraft with fly-by-wire are limited in ways that change under different conditions. Parameters such as angle of attack, bank angle, roll rate, and engine power, among other things, are monitored, managed systemically, and limited, no matter what the pilot does with the stick. This has the effect of making it impossible to stall* the airplane, and it simplifies the number of things a pilot needs to remember to do in response to certain emergency situations.
* A "stall" in an airplane is not what the layman might think—the engine doesn't quit. This technical term means that the aircraft wing angle of attack is so steep that the wing ceases to produce lift (because airflow is disrupted over the upper surface). This is dangerous because it can lead to uncontrolled descent, especially a spin. The angle of attack in a stall is always the same for a given aircraft, but the stall speed varies as a function of many variables including weight, density altitude, and bank angle.
Boeing held steadfastly to the mechanical control model until the 777, which is Boeing's first fly-by-wire passenger aircraft. The reason for being late to adopt this technology was explained to me by a retired Boeing senior engineering director, "You just don't know what's going to happen to those electrons between the front and the back of the plane. I like a direct connection better."
For a time, the old view seemed the safest as Airbus worked out its fly-by-wire bugs in a very public fashion at an air show in France in 1988 with the crash of an A320 into a forest while performing a low-altitude, low-speed fly-by. There were a few other related crashes*, but in recent years the technology has been improved, proven in commercial service, and incorporated into all Airbus aircraft developed after the A320.
* Engineering bulletins issued by Airbus to operators indicated that certain aspects of the many interrelated inputs and controls were still being worked out at the time but apparently had not been applied to that specific airplane. Another A320 crashed in Bangalore, India, in 1990 and a third in France in 1992 fueled discussion about whether the new technology was ready for commercial service.
When Boeing adopted fly-by-wire on the 777, it came with a big difference—the ability for the pilot to override the computer's limits. Boeing argues that the pilot should be the ultimate judge of whether an emergency requires going beyond standard operating and safety parameters. As an example, a Boeing spokesman5 cited a China Air 747 incident in 1985 where the crew recovered from an out-of-control dive with a recovery that stressed the airplane at up to 4g's, something that Airbus fly-by-wire would limit to 2.5g's, perhaps limiting its capability to recover from some unusual attitudes.
Those who advocate unlimited pilot control believe that fly-by-wire limits are akin to saving the airplane from overstress but crashing it in the process. Advocates of fully integrated system control tell the old joke that describes the computer-controlled airplane of the future as having a seat in the cockpit for a pilot and a dog. The pilot's job is to the feed the dog, and the dog's job is to bite the pilot if he tries to touch anything.
Can technology save us from our own mistakes? Yes and no. Technology can improve the odds when we understand the range of possible actions of something like an aircraft or another technically controlled machine or system. But most businesses have many dimensions, and not all have accepted preprogrammed responses, so while technology may help, it is unlikely to stop business mistakes.
Debates about the appropriate use of technology are constant. In recent months, these have included issues such as whether the New York Stock Exchange should be replaced with an electronic exchange and whether the electric power grid in the United States can be improved with more technology. These and other examples involve complex systems of human, economic, and technical interaction with a range of parameters under normal conditions. Yet all have the potential to spiral out of control when there are multiple mistakes or unusual situations that were not anticipated and built into operating parameters and designs.
This is the conundrum that surrounds multiple mistakes. We can anticipate many, but not all, mistakes that people or systems will induce in business or the operation of complex machines. If we can anticipate mistakes, should we train people to avoid the circumstances or build technology-based systems that prevent those things from happening? If we build programmed error-control systems, will we induce more mistakes or prevent recovery from mistakes we did not anticipate?
We can use technology to improve business processes like the supply chain, but computers cannot decide how you will identify and design new products that go into that supply chain or where and how they will be manufactured. This is where physical systems and businesses diverge. Business systems still require judgment, thus we need to continue to refine and improve the quality of judgment and decision-making abilities of individuals operating businesses.