The Most Human Human

by Brian Christian

  –JOSEPH WEIZENBAUM

  The term method itself is problematic because it suggests the notion of repetition and predictability—a method that anyone can apply. Method implies also mastery and closure, both of which are detrimental to invention.

  –JOSUÈ HARARI AND DAVID BELL

  Pure technique, Weizenbaum calls it. This is, to my mind, the crucial distinction. “Man vs. machine” or “wetware vs. hardware” or “carbon vs. silicon”–type rhetoric obscures what I think is the crucial distinction, which is between method and method’s opposite: which I would define as “judgment,” “discovery,”3 “figuring out,” and, an idea that we’ll explore in greater detail in a couple pages, “site-specificity.” We are replacing people not with machines, nor with computers, so much as with method. And whether it’s humans or computers carrying that method out feels secondary. (The earliest games of computer chess were played without computers. Alan Turing would play games of “paper chess” by calculating, by hand, with a pencil and pad, a move-selection algorithm he’d written. Programming this procedure into a computer merely makes the process go faster.) What we are fighting for, in the twenty-first century, is the continued existence of conclusions not already foregone—the continued relevance of judgment and discovery and figuring out, and the ability to continue to exercise them.

  Reacting Locally

  “Rockstar environments develop out of trust, autonomy, and responsibility,” write programmers and business authors Jason Fried and David Heinemeier Hansson. “When everything constantly needs approval, you create a culture of nonthinkers.”

  Fellow business author Timothy Ferriss concurs. He refers to micromanagement as “empowerment failure,” and cites an example from his own experience. He’d outsourced the customer service at his company to a group of outside representatives instead of handling it himself, but even so, he couldn’t keep up with the volume of issues coming in. The reps kept asking him questions: Should we give this guy a refund? What do we do if a customer says such and such? There were too many different cases to make setting any kind of procedure in place feasible, and besides, Ferriss didn’t have the experience necessary to decide what to do in every case. Meanwhile, questions kept pouring in faster than he could deal with them. All of a sudden he had an epiphany. You know who did have the experience and could deal with all these different unpredictable situations? The answer was shamefully obvious: “the outsourced reps themselves.”

  Instead of writing them a “manual,” as he’d originally planned, he sent an email that said, simply, “Don’t ask me for permission. Do what you think is right.” The unbearable stream of emails from the reps to Ferriss dried up overnight; meanwhile, customer service at the company improved dramatically. “It’s amazing,” he says, “how someone’s IQ seems to double as soon as you give them responsibility and indicate that you trust them.” And, as far too many can attest, how it halves when you take that responsibility and trust away.

  Here in America, our legal system treats corporations, by and large, as if they were people. As odd as this is, “corporation” is etymologically bodily, and bodily metaphors for human organizations abound just about everywhere. There’s a great moment in the British series The Office when David Brent waxes to his superior about how he can’t bring himself to make any layoffs because the company is “one big animal. The guys upstairs on the phones, they’re like the mouth. The guys down here [in the warehouse], the hands.” His boss, Jennifer, is senior management: the “brain,” presumably. The punch line of the scene is, naturally, that David—who of course is the most layoff-worthy yet is the one in charge of the layoffs—can’t figure out what organ he is, or what role he plays in the organization.

  But there’s a deeper point worth observing here too, which is that we create a caste system at our companies that mimics the caste system we create with respect to our own bodies and selves. My hands are mine, we say, but my brain is me. This fits in nicely with our sense of an inner homunculus pulling the levers and operating our body from a control room behind our eyeballs. It fits in nicely with Aristotle’s notion that thinking is the most human thing we can do. And so we compensate accordingly.

  I almost wonder if micromanagement comes from the same over-biasing of deliberate conscious awareness that led, both to and out of, the Turing machine model of computation underlying all of our computers today. Aware of everything, acting logically, from the top down, step-by-step. But bodies and brains are, of course, not like that at all.

  Micromanagement and out-of-control executive compensation are odd in a way that dovetails precisely with what’s odd about our rationalist, disembodied, brain-in-a-vat ideas about ourselves. When I fight off a disease bent on my cellular destruction, when I marvelously distribute energy and collect waste with astonishing alacrity even in my most seemingly fatigued moments, when I slip on ice and gyrate crazily but do not fall, when I unconsciously counter-steer my way into a sharp bicycle turn, taking advantage of physics I do not understand using a technique I am not even aware of using, when I somehow catch the dropped oranges before I know I’ve dropped them, when my wounds heal in my ignorance, I realize how much bigger I am than I think I am. And how much more important, nine times out of ten, those lower-level processes are to my overall well-being than the higher-level ones that tend to be the ones getting me bent out of shape or making me feel disappointed or proud.

  Software development gurus Andy Hunt and Dave Thomas make the point that with a certain latitude of freedom and autonomy, a greater sense of ownership of a project emerges, as well as a sense of artistry; as they note, the stonemasons who helped build cathedrals were far from drones—they were “seriously high quality craftsmen.”

  The idea of artistic freedom is important because it promotes quality. As an example, suppose you’re carving a gargoyle up in the corner of this building. The original spec either says nothing or says you’re making a straight on gargoyle just like these others. But you notice something because you’re right there on the ground. You realize, “Oh look, if I curve the gargoyle’s mouth this way, the rain would come down here and go there. That would be better.” You’re better able to react locally to conditions the designers probably didn’t know about, didn’t foresee, had no knowledge of. If you’re in charge of that gargoyle, you can do something about that, and make a better overall end product.

  I tend to think about large projects and companies not as pyramidal/hierarchical, per se, so much as fractal. The level of decision making and artistry should be the same at every level of scale.

  The corporate isn’t always a great example of this. The corporeal is, as is another kind of organization with the body in its etymology—the U.S. Marine Corps. Consider this, from their classic handbook Warfighting:

  Subordinate commanders must make decisions on their own initiative, based on their understanding of their senior’s intent, rather than passing information up the chain of command and waiting for the decision to be passed down. Further, a competent subordinate commander who is at the point of decision will naturally better appreciate the true situation than a senior commander some distance removed. Individual initiative and responsibility are of paramount importance.

  In some sense this question of management style, of individual responsibility and agency, cuts not only across traditional divisions of “blue”- and “white”-collar work, but also between “skilled” and “unskilled” work. A formulaic mental process rigidly repeated time and again is not so different from a physical process repeated in this way. (That is to say, there’s such a thing as thinking unthinkingly.) Likewise, a complex or sophisticated or learnèd process repeated time and again is not so different from a simple process repeated. More important than either of these distinctions, arguably, is the question of how much reacting locally or site-specificity, how much freshness in approach the job requires—or permits.

  In March 2010, National Public Radio’s This American Life did a segment on the joint General Moto
rs and Toyota plant NUMMI. One of the biggest differences, it turned out, between the two companies was that at Toyota, “when a worker makes a suggestion that saves money, he gets a bonus of a few hundred dollars or so. Everyone’s expected to be looking for ways to improve the production process. All the time. This is the Japanese concept of kaizen, continuous improvement.” One American GM worker was part of a group that traveled to Japan to try building cars on the Toyota assembly line, and the difference in the experience stunned him:

  I can’t remember any time in my working life where anybody asked for my ideas to solve the problem. And they literally want to know. And when I tell them, they listen. And then suddenly they disappear and somebody comes back with the tool that I just described. It’s built, and they say, “Try this.”

  One of the results of this kind of participation is that “IQ-doubling” effect that Ferriss describes. You’re not just doing something; you’re doing that very human thing of simultaneously stepping back and considering the process itself. Another effect: pride. NUMMI’s union leader, Bruce Lee, said he’d never felt about the cars he built like he did once he was participating in the process: “Oh, I was so proud of ’em you can’t believe.”

  A Robot Will Be Doing Your Job

  For the many, there is a hardly concealed discontent. The blue-collar blues is no more bitterly sung than the white-collar moan. “I’m a machine,” says the spot-welder. “I’m caged,” says the bank teller, and echoes the hotel clerk. “I’m a mule,” says the steelworker. “A monkey can do what I do,” says the receptionist. “I’m less than a farm implement,” says the migrant worker. “I’m an object,” says the high-fashion model. Blue collar and white call upon the identical phrase: “I’m a robot.”

  –STUDS TERKEL

  The notion of computer therapists of course raises one of the major things that people think of when AI comes to mind: losing their jobs. Automation and mechanization have been reshaping the job market for several centuries at this point, and whether these changes have been positive or negative is a contentious issue. One side argues that machines take human jobs away; the other side argues that increased mechanization has resulted in economic efficiency that raises the standard of living for all, and that has released humans from a number of unpleasant tasks. The corollary to the “advance” of technology seems to be that familiar human “retreat,” for better and for worse.

  We call a present-day technophobe a “Luddite,” which comes from a group of British workers who from 1811 to 1812 protested the mechanization of the textile industry by sabotaging mechanical looms:4 this debate has been going on—in words and deeds—for centuries. But software, and particularly AI, changes this debate profoundly—because suddenly we see the mechanization of mental work. As Matthew Crawford argues in the 2009 book Shop Class as Soulcraft, “The new frontier of capitalism lies in doing to office work what was previously done to factory work: draining it of its cognitive elements.”

  I would also like to note something, though, about the process by which jobs once performed by humans get taken over by machines, namely, that there’s a crucial intermediate phase to that process: where humans do the job mechanically.

  Note that the “blue collar and white” workers complaining about their robotic work environments in Terkel’s 1974 book Working are bemoaning not jobs they’ve lost, but the jobs they have.

  This “draining” of the job to “robotic” behavior happens in many cases long before the technology to automate those jobs exists. Ergo, it must be due to capitalist rather than technological pressures. Once the jobs have been “mechanized” in this way, the much later process by which those jobs actually get taken over by machines (or, soon, AIs) seems like a perfectly sensible response, and, by that point, perhaps a relief. To my mind, the troubling and tragic part of the equation is the first half—the reduction of a “human” job to a “mechanical” one—and less so the second. So fears over AI would seem to miss the point.

  Micromanagement; the kaizen-less assembly line; the over-standardization of brittle procedures and protocols … these problems are precisely the same problem, and pose precisely the same danger, as does AI. In all four cases, a robot will be doing your job. The only difference is that in the first three, the robot will be you.

  “I Can’t Communicate”

  We’re in the midst of an interesting moment for AI chatbots, which are finally starting to show commercial promise. Just recently, the Alaska Airlines website wanted me to chat with “Jenn” instead of using their customer service telephone number (I declined), and that’s just the most recent example of many.5 But before chatbots, there was, of course, the much-loathed automated telephone menu systems. And before that, human operators were behaving like chatbot automatons. Operator Heather Lamb, for instance, says in Working, “There are about seven or eight phrases that you use and that’s it. ‘Good morning, may I help you?’ ‘Operator, may I help you?’ ‘Good afternoon.’ ‘Good evening.’ ‘What number did you want?’ ‘Would you repeat that again?’ ‘I have a collect call for you from so-and-so, will you accept the charge?’ ‘It’ll be a dollar twenty cents.’ That’s all you can say. A big thing is not to talk with a customer … I’m a communications person but I can’t communicate.”

  I’ve called directory assistance a handful of times in recent years: the folks are as monosyllabic as possible, and brusque to the point of inhumanity. If my interaction with them is “human” in any way, it is only in the way that stepping on a stranger’s toes on the bus and getting scowled at is “human.” It’s not their fault, of course—they’re being forced to act like bots. In this particular case, I’ll take the robot: at least I don’t have to feel like a nuisance.

  Now, if the people at directory assistance lived nearby, and could offer useful suggestions like “Oh, do you mean Dave’s downtown or the David’s up on Fifteenth?” or “But actually, if you’re looking for a good steak my personal recommendation would be …,” that would be a different story entirely. But they don’t live near where you’re calling (scaling), they aren’t given the time to engage with you (efficiency), and they can’t really deviate from the script (pure technique).

  Just today, I call to activate my new credit card and end up on the phone for a good ten minutes: the woman is getting snowed on in northern Colorado, wishing for milder weather, and I’m getting rained on in Seattle, wishing for a wintrier winter. Being from the Jersey shore, I’ve grown up accustomed to snowy winters and muggy summers. Sometimes I love the Northwest’s temperateness; sometimes I miss the Northeast’s intensity. The shore, she says, wow, I’ve never seen the ocean … And on from there. My roommate, passing through the living room, assumes it’s an old friend on the line. Eventually the card is activated, and I snip the old one and wish her the best.

  Maybe it’s not until we experience machines that we appreciate the human. As film critic Pauline Kael put it, “Trash has given us an appetite for art.” The inhuman has not only given us an appetite for the human; it’s teaching us what it is.

  Maggot Therapy

  It’s clear from all of this that AI is not really the enemy. In fact, it may be that AI is what extricates us from this process—and what identifies it. Friends of mine who work in software talk about how a component of their job often involves working directly on problems while simultaneously developing automated tools to work on those problems. Are they writing themselves out of a job? No, the consensus seems to be that they move on to progressively harder, subtler, and more complex problems, problems that demand more thought and judgment. They make their jobs, in other words, more human.

  Likewise, friends of mine not in software—in PR, marketing, you name It—are increasingly saying to me: “Can you teach me how to program? The more you talk about scripting … I’m pretty sure I can automate half my job.” In almost all cases they are right.

  I think, quite earnestly, that all high school students should be taught how to program. It will give our next ge
neration a well-deserved indignation at the repetitiveness and rule-bound-ness of some of the things they’ll be asked to do. And it will also give them the solution.

  You can almost think of the rise of AI not as an infection or cancer of the job market—the disease is efficiency—but as a kind of maggot therapy: it consumes only those portions that are no longer human, restoring us to health.

  Art Cannot Be Scaled

  Aretê … implies a contempt for efficiency—or rather a much higher idea of efficiency, an efficiency which exists not in one department of life but in life itself.

  –H. D. F. KITTO, THE GREEKS, QUOTED

  IN ROBERT PIRSIG’S ZEN AND THE ART

  OF MOTORCYCLE MAINTENANCE

  A farm equipment worker in Moline complains that the careless worker who turns out more that is bad is better regarded than the careful craftsman who turns out less that is good. The first is an ally of the Gross National Product. The other is a threat to it, a kook—and the sooner he is penalized the better. Why, in these circumstances, should a man work with care? Pride does indeed precede the fall.