The Design of Future Things

by Don Norman

  The problem is that a wobbly steering wheel would make the driver think there was something the matter with the car. Not only would this send the wrong message, but it would never be acceptable to the automobile manufacturers. When I mentioned this to an engineering group at one major automobile company, the response was nervous laughter. “Why would we want to produce a product that sometimes felt as if it were failing to work properly?” they asked. Good point. Instead of making the car appear to be dangerous, however, we could make it appear that the environment was dangerous.

  Picture someone driving over an old dirt highway, with deep ruts that capture the car, moving it in unpredictable ways: in such a situation, we wouldn’t blame the car; we would blame the ruts in the road. Or consider driving through thick, heavy mud, which makes the automobile behave sluggishly and unable to respond with agility. Or, if driving on ice so that the car continually skidded, we would slow down and take caution but, once again, put the blame on the road. Finally, consider driving on a clear, sunny day, on a modern highway with no other traffic in sight. The car can respond with agility and promptness: now we would believe the responses to be entirely due to the car itself.

  All these environmental variables would have a desirable impact upon the driver’s responses, but in a manner attributed to the environment, not the automobile. This would naturally induce just the correct behavior: the more dangerous something appears to be, the more care the person in charge will exert.

  Why do we need this? Because the modern automobile has become too comfortable. Between the effective shock absorbers and ride-control systems, effective noise reduction, and minimization of road feel and vibration from the interior of the automobile, the driver is no longer in direct touch with the environment. As a result, artificial enhancement is required, the better to let the driver know the environmental conditions.

  Note that I am not in favor of actually making things more dangerous. The goal is that by providing appropriate feedback, the driver will drive more safely. Of course, we should continue to enhance true physical safety. We know that completely automatic systems have already been proven effective: for example, antiskid brakes and stability controls, smoke alarms, helmets for bicyclists and skateboarders and skiers, as well as shields and guards for machinery, are all important. Yet, these automatic systems have limited effectiveness. If drivers drove more safely in the first place, then the automatic systems would be a lot more effective when the unexpected did occur.

  These ideas are controversial. Even I am not completely convinced they would work. Human nature being what it is, people are very apt to do just the opposite of my predictions, ignoring the apparent slipperiness of the road by thinking, “Oh, the road isn’t really slippery. This is just the car trying to slow me down.” But what if the road really is slippery? Furthermore, would you buy a car or tool that deliberately frightened you? Bad marketing strategy. Bad idea.

  Still, there is truth in the phenomenon. Today, we are too comfortable, too insulated from the dangers inherent in the world, inherent in the operation of complex, powerful machinery. If motorcycles and automobiles, machinery and drugs seemed as dangerous as they really are, perhaps people would modify their behavior appropriately. When everything is soundproofed, cushioned, and sanitized, we are no longer aware of the real hazards. That is why we need to bring back the truthful depiction of the danger.

  Responsive Automation

  Power-assisted devices, such as brakes and steering, are relatively primitive examples of a natural collaboration between person and machine. With modern electronics, much more collaboration is possible. Consider the “Cobot,” or “Collaborative Robot,” invented by Professors Ed Colgate and Michael Peshkin at their Laboratory for Intelligent Mechanical Systems at Northwestern University. Cobots are another excellent example of a natural interaction between a person and a machine, akin to the interaction between a horse and rider. When I asked Peshkin to describe Cobots, here is how he responded:

  The smartest things are those that complement human intelligence, rather than try to supersede it. Much like the smartest teacher.

  The point of the Cobot is shared control and shared intelligence between the person and the device. The robot does what it does well, and the person what people do well.

  Our first applications are in material handling, in automobile assembly and warehousing. Here the Cobot provides smooth guiding surfaces that the human can use to help move a payload quickly and accurately and more ergonomically. When the payload is not in contact with a virtual surface, the human can move the payload at will, applying vision, dexterity, problem solving skills. And when necessary, push it up against a guiding surface and swoop along it.

  The Cobot provides an excellent example of human-machine symbiosis because, as far as the people who use it are concerned, they are simply lifting and moving objects as they normally would. The only difference is that these objects might be extremely heavy; yet, only small lifting and guiding forces are required. The system amplifies force: people only need to exert small, comfortable amounts of force, and the system supplies whatever else is required. The people feel as if they are in complete control and may even be unaware that they are being assisted by mechanical means. For example, one application of Cobot technology helps automobile assembly line workers manipulate automobile engines. Normally, heavy objects, such as automobile engines, are lifted by overhead hoists that must be controlled or by intelligent hoists that try to do the task by themselves automatically. With the Cobot, workers simply loop a chain and hook around the engine and lift up. The engine is far heavier than one person can lift, let alone with one hand, but the Cobot, sensing the lifting action, supplies the lifting force that is required. When workers want the engine to move, or to rotate, or to be lowered again, they simply lift, rotate, push, or press down; the Cobot senses the forces and amplifies them appropriately for the task. The result is a perfect collaboration. The workers do not think of themselves as using a machine: they just think of themselves as moving the engine.

  Cobots can be a lot more sophisticated, too. For example, if it is important not to move the engine in some directions, or if it is important to carry the engine along a well-described path, the Cobot control system can define virtual walls and paths so that when the user tries to go against the wall or deviate from the path, the device “pushes back,” resisting the attempt, but still in a nice, natural manner. In fact, the worker can use this artificial wall as an aid, pushing the engine all the way to the side until it hits the “wall,” then sliding the engine along the wall. Artificially induced limits of this sort feel amazingly natural. The machine doesn’t seem to be forcing a behavior: it feels like a physical wall, and so it is natural either to avoid it or perhaps to use it as an aid in maintaining a straight path by deliberately contacting it and sliding along it. Here is how the originators of Cobots described this possibility:

  One of the most exciting capabilities . . . is the implementation of programmable constraint. For example, hard walls which constrain motion to useful directions can dramatically improve performance in tasks such as remote peg-in-hole insertion. Another example . . . [is] the “Magic Mouse,” a computer interface device which can constrain an operator’s hand to useful directions . . . to avoid, for instance, “slipping off” a pull-down menu. A third example is a robotic surgery system in which a robot positions a guide for a tool held by a surgeon, and a fourth is automobile assembly in which programmed constraints can help an operator navigate large components (e.g., instrument panels, spare tires, seats, doors) into place without collisions.

  Cobots are part of a family of power-assisted systems. One other example is that of a powered exoskeleton, a type of suit or mechanical skeleton put on over the body that, just as the Cobot, senses the person’s movements and amplifies them to whatever degree necessary. Exoskeletons are still more of a concept than reality, but proponents of these future things foresee their use in construction, firefighting, and oth
er dangerous environments, enabling a person to lift heavy loads and jump long distances and heights. They could also be beneficial to medical treatments, allowing impaired patients to have normal strength, while also providing rehabilitation training by gradually increasing the amount of force the patient is required to supply, thereby guiding the rehabilitation process. Much like the use of the horse metaphor for automobile control, which can vary between loose rein and tight rein, medical rehabilitation exoskeletons could vary between having the patient in charge (tight rein) and having the robot in charge (loose rein).

  Another example of natural interaction is the Segway Personal Transporter, a two-wheeled personal transportation system. The transporter provides a powerful example of how intelligent design can yield a vehicle that is a wonderful symbiosis of machine+person. The transporter provides behavioral control and the human, high-level reflective guidance. Stand on the transporter, and it automatically balances you and it, together. Lean forward, and it moves forward; lean back, and it stops. Similarly, to turn, just lean in the correct direction. It’s easier to use than a bicycle, and the interaction feels natural. The Segway transporter isn’t for everyone, however, just as a horse isn’t for everyone. It requires some degree of skill and attentiveness.

  Contrast the natural interaction of horse and rider, person and Cobot, or person and the Segway Transporter with the more rigid interaction between a person and the automatic flight control of an airplane or even the cruise control of an automobile. In the latter, the designers assume that you deliberately set the control, turn it on, and then have nothing more to do—until it fails, that is, when you are suddenly asked to recover from whatever problem has overwhelmed the automation.

  FIGURE 3.4

  The Segway® Personal Transporter. A kind of collaborative robot, where control is done by leaning in one direction or another. Naturally, easily, both human and the transporter form a symbiotic unit.

  (Photo used with permission of Segway Media.)

  The examples of natural, responsive interaction discussed in this section illustrate a natural application of machine intelligence and collaborative power to provide a true machine+person symbiosis—human-machine interaction at its best.

  CHAPTER FOUR

  Servants of Our

  Machines

  MOTORIST TRAPPED IN TRAFFIC

  CIRCLE 14 HOURS

  April 1. Hampstead, MA. Motorist Peter Newone said he felt as if a nightmare had just ended. Newone, 53, was driving his newly purchased luxury car when he entered the traffic circle in the city center around 9 a.m. yesterday, Friday. The car was equipped with the latest safety features, including a new feature called Lane Keeping. “It just wouldn’t let me get out of the circle,” said Newone. “I was in the inner-most lane, and every time I tried to get out, the steering wheel refused to budge and a voice kept saying over and over, ‘warning, right lane is occupied.’ I was there until 11 at night, when it finally let me out,” Newone said from his hospital bed, his voice still shaky. “I managed to get out of the circle and to the side of the road, and then I don’t remember what happened.”

  Police say they found Newone collapsed in his car, incoherent. He was taken to the Memorial Hospital for observation and diagnosed with extreme shock and dehydration. He was released early this morning.

  A representative of the automobile company said that they could not explain this behavior. “Our cars are very carefully tested,” said Mr. Namron, “and this feature has been most thoroughly vetted by our technicians. It is an essential safety feature and it is designed so that it never exerts more than 80% of the torque required, so the driver can always overrule the system. We designed it that way as a safety precaution. We grieve for Mr. Newone, but we are asking our physicians to do their own evaluation of his condition.”

  Police say they have never heard of a similar situation. Mr. Newone evidently encountered a rare occurrence of continual traffic at that location: there was a special ceremony in the local school system which kept traffic high all day, and then there was an unusual combination of sports events, a football game, and then a late concert, so traffic was unusually heavy all day and evening. Attempts to get statements from relevant government officials were unsuccessful. The National Transportation Safety Board, which is supposed to investigate all unusual automobile incidents, says that this is not officially an accident, so it does not fit into their domain. Federal and state transportation officials were not available for comment.

  Cautious cars, cantankerous kitchens, demanding devices. Cautious cars? We already have them, cautious and sometimes frightened. Cantankerous kitchens? Not yet, but they are coming. Demanding devices? Oh, yes, our products are getting smarter, more intelligent, and more demanding, or, if you like, bossy. This trend brings with it many special problems and unexplored areas of applied psychology. In particular, our devices are now part of a human-machine social ecosystem, and therefore they need social graces, superior communicative skills, and even emotions—machine emotions, to be sure, but emotions nonetheless.

  If you think that the technologies in your home are too complex, too difficult to use, just wait until you see what the next generation brings: bossy, demanding technologies, technologies that not only take control of your life but blame you for their shortcomings. It is tempting to fill this book with horror stories, real ones that are happening today plus imagined ones that could conceivably come about if current trends continue, such as the imaginary story of Mr. Newone.

  Consider poor Mr. Newone, stuck in the traffic circle for fourteen hours. Could this really happen? The only real clue that the story isn’t true is the date, April 1, because I wrote this story specifically for the yearly April Fool’s edition of the RISKS digest, an electronic newsletter devoted to the study of accidents and errors in the world of high technology. The technologies described in the article are real, already available on commercially sold automobiles. In theory, just as the spokesperson in the story says, they only provide 80 percent of the torque required to stay in the lane, so Mr. Newone presumably could have overcome that force easily. Suppose he were unusually timid, however, and as soon as he felt the resisting force on the steering wheel, he would immediately give in. Or what if there were some error in the mechanics, electronics, or programming of the system, causing 100 percent of the force to be deployed, not 80 percent. Could that happen? Who knows, but the fact that it is so plausible is worrisome.

  We Have Become the Tools of Our Tools

  But lo! men have become the tools of their tools.

  —Henry Thoreau, Walden

  When Henry Thoreau wrote that “men have become the tools of their tools,” he was referring to the relatively simple tools of the 1850s, such as the axe, farming implements, and carpentry. Even in his time, however, tools defined people’s lives. “I see young men, my townsmen, whose misfortune it is to have inherited farms, houses, barns, cattle, and farming tools; for these are more easily acquired than got rid of.” Today, we complain about the maintenance all our technology requires, for it seems never ending. Thoreau would have sympathized, for even in 1854 he compared the daily toil of his neighbors unfavorably to the twelve labors of Hercules: “The twelve labors of Hercules were trifling in comparison with those which my neighbors have undertaken; for they were only twelve, and had an end.”

  Today, I would rephrase Thoreau’s lament as “People have become slaves to their technology, servants of their tools.” The sentiment is the same. And not only must we serve our tools, faithfully using them throughout the day, maintaining them, polishing them, comforting them, but we also blithely follow their prescriptions, even when they lead us to disaster.

  It’s too late to go back: we can no longer live without the tools of technology. Technology is often blamed as the culprit: “technology is confusing and frustrating,” goes the standard cry. Yet, the complaint is misguided: most of our technology works well, including the tool Thoreau was using to write his complaint. For that matter,
Thoreau himself was a technologist, a maker of tools, for he helped improve the technology in the manufacture of pencils for his family’s pencil business. Yes, a pencil is a technology.

  Tech·nol·o·gy (noun): New stuff that doesn’t work very well or that works in mysterious, unknown ways.

  In the common vernacular, the word “technology” is mostly applied to the new things in our life, especially those that are exotic or weird, mysterious or intimidating. Being impressive helps. A rocket ship, surgical robots, the internet—that’s technology. But a paper and pencil? Clothes? Cooking utensils? Contrary to the folk definition, the term technology really refers to any systematic application of knowledge to fashion the artifacts, materials, and procedures of our lives. It applies to any artificial tool or method. So, our clothing is the result of technology, as is our written language, much of our culture; even music and art can be thought of as either technologies or products that could not exist without the technologies of musical instruments, drawing surfaces, paints, brushes, pencils, and other tools of the artists and musicians.