Mindless: Why Smarter Machines Are Making Dumber Humans
by Simon Head
Basic Books, 230 pp., $26.99
The entire thesis of Simon Head’s arresting new book is contained in the subtitle. It goes all the way back to Adam Smith’s telling observation that the division of labor in a pin factory, while doing wonders for productivity (output per worker), would make workers as “stupid and ignorant as it is possible for a human creature to be.” This was because no worker needed to know how to make a pin, only how to do his part in the process of making a pin. Artisan production was on the point of becoming industrial production; industrial production would destroy work skills.
By the start of the twentieth century, and after the Industrial Revolution, Smith’s pin factory had become Henry Ford’s Rouge plant. “What was worked out at Ford,” wrote Charles Sorensen in his memoir on his years with the car manufacturer, “was the practice of moving the work from one worker to another until it became a complete unit, then arranging the flow of these units at the right time and the right place to a moving final assembly line from which came a finished product.” The finished product was of course the Ford motor car. It took 1.33 man hours (as against the previous time of 12.5 hours) to produce a Model T Ford, and they came off the assembly line every three minutes.
With fewer workers needed to produce each car, wages per worker could go up and hours of work could be reduced. But because each car was cheaper to produce, the volume of sales could be hugely expanded, so the number of workers newly employed in manufacturing motor cars far exceeded those displaced. Contrary to the fear of the Luddites, machinery was adding to employment, not subtracting from it. But the Luddites—originally skilled workers in the Midlands and North of England who smashed textile machinery between 1811 and 1817—protested not just against loss of jobs and wages, but of skills and communities.
The point of the assembly line was that it must be kept moving, the faster the better. Any breakdown brought the production process to a juddering halt. Continuity of production required a high level of managerial control over work practices, in other words, “scientific management.” This was the invention of Frederick Winslow Taylor. Fordism grew up with Taylorism. As Simon Head writes, “in political and sociological terms, Taylorism can be seen as the division of labor pushed to its logical extreme, with a consequent de-skilling of the worker and dehumanisation of the workers and the workplace.” Taylor’s disciple, William Henry Leffingwell, began applying the methods of scientific management to the service sector from the 1920s onward, and today it is almost ubiquitous.
“Scientific management” is Simon Head’s point of entry—and protest—in his fine book. Head is a journalist turned academic who has specialized in writing about the social impact of technology. In The New Ruthless Economy (2003), he analyzed the practice of call centers, showing how digitalized scripts required of their operators robot-like behavior. In his latest book he claims that computer programming is now applied to all the principal sectors of the manufacturing and service economy.
The upshot is that networked computers, with monitoring software attached, have hugely expanded “the power to manage the affairs of giant global corporations and…micromanage the work of their single employees or teams of employees.” Their possibilities have spawned “Computer Business Systems” (CBS) which have colonized much of the service sector.
The tendency of CBS, Head argues, is to discourage intuition and judgment in a large population, except for a tiny class of highly paid engineers and managers, who are needed to activate and control the automated systems. What Head calls “digital managerialism” achieves this by transforming the objects of management into “electronic representations” of human beings, “the numbers, coded words, cones, squares, and triangles that represent us on [the] digital screens [of managers].” Such electronic representations have been applied increasingly to middle management, who, deprived of their traditional oversight functions, are themselves subject to the intrusive monitoring of time and performance they had exercised over their subordinates.
The three interrelated elements of CBS are: computer networks (the Internet) linking “the work station of every employee or group of employees within an organization to that of every other”; “data warehouses,” containing “the gigantic quantities of information” needed to monitor the actions of employees “in real time” and control them “in line with matrices established by management”; and “expert systems that mimic human intelligence in performing the cognitive tasks” integral to personal services.
Head’s account pivots on the distinction between “process” and “practice.” Process refers to “a series of operations and how they relate to one another.” Practice refers to the “accumulation of tacit knowledge and skill” that employees bring to their tasks. In the automated systems, “‘process’…pushes ‘practice’ aside.”
The great strength of Head’s approach is that he deconstructs and demystifies for the nonexpert reader the pseudoscientific, abstract, jargonized language of management studies, in order to reveal the dispiriting reality it obscures. The aim of all control systems is to control human behavior, including the way we think. Priests and political leaders have long used religion and ideology for this purpose, since it economizes on the use of force and terror. But it is only in the last hundred years or so that the attempt to control behavior by controlling the mind has achieved scientific status, largely through the explosion of calculating power that computers have made possible. In one of his many fascinating chapters, Head shows how CBS originated in the needs of the military for battlefield control, before they were applied to the needs of business.
Unlike the machine assembly line for Ford cars, the human assembly lines in giant retail organizations like Walmart and Amazon pose special problems. The stacking and retrieving of customers’ orders requires the attention of a “panoptic monitoring regime to pick up on…human waywardness [on the part of the employees] and correct it without delay.” The model is that of Jeremy Bentham’s Panopticon, the circular prison he designed with an inspection tower at its center, where a single watchman could observe the inmates without them being able to tell if they are being watched. Bentham himself thought of the Panopticon as an unprecedented way of obtaining power of mind over mind. What has made computerized business systems universally applicable is the joining of Taylorian scientific management (breaking down jobs into small tasks) with the panoptical control made possible by digital technology.
Readers of The New York Review, most of whom probably have more control over their time than employees of Walmart, may be inclined to dismiss CBS manuals as the fantasies of the impotent. And such readers are right, up to a point: human beings are notoriously recalcitrant to attempts to hammer them into the required shape. Most political dystopias like 1984 or Brave New World succumb to some outbreak of the human spirit. And while Head brilliantly translates ideas remote from the experience of most people into everyday language, he is too inclined to take the claims of the engineering manuals for reality. He would have done better in some cases to talk to the designers of these systems and ask them what they were really hoping to achieve; and to try to get a closer feel for life in automated distribution systems by working in one of them undercover, as Carole Cadwalladr did at Amazon.
What he does show convincingly is that these control systems have permeated deeply into the service sector, increasing productivity in activities that were assumed to be relatively immune to them. He does not mention the thesis of the economist William Baumol, who argued that there exists a class of goods whose production cannot be automated, and whose cost therefore is bound to rise relative to those goods that can. The examples Baumol gave were from the performing arts, but his idea was generalized to include all those goods and services whose value depends on person-to-person contact. What Head shows is that the class of such “Baumol goods” may be shrinking. Since the service sector now makes up 70 to 80 percent of Western economies, it is right for us to take notice.
In a powerful chapter, Head analyzes the methods used by Walmart and Amazon to squeeze ever more production out of their workers, through pervasive control of the human conveyor belt—warehouse to shop for Walmart, warehouse to customer for Amazon. “Speeding up” is a constant preoccupation of the senior managment in a conveyor belt system, whether of humans or machines, since the faster the speed, the lower the per unit cost. Research at Foxconn’s factory in China showed that “if workers can finish their quota the target will be increased day by day until the capacity of the workers is maximised.” All this is meant to be in the service of customers. But as Head pertinently asks:
Should these marginal benefits to customers really be purchased at the price of a system that treats employees as untrustworthy human robots and relies on intimidation to push them to the limit, while denying them the rewards of their own increased efficiency?
Digital control systems have now penetrated even into those parts of the service sector that require “cognitive functions,” i.e., where the objects of production are not consumables, but “the treatment of sick patients, the transactions between teachers and pupils, or the decisions to hire and fire employees.” More and more important for such production are engineers of the mind and the emotions in the form of “human resources” and “customer relations” experts. For example, it can be calculated how much smiling flight attendants need to do to make passengers feel they are being sufficiently pampered.
But humans are recalcitrant. Flight attendants at Cathay Pacific in 2012 responded to attempts to speed up their work by threatening to go on “smile strikes.” Attempts to create a happy demeanor by encouraging workers to think of pleasant past experiences led to daydreaming that hindered efficiency. In the case of financial products like CDOs—collateral default obligations—assembled from often shaky mortgages and sold on by banks, the automating of the “cognitive functions” with digital scripts at points along the production line helped cause the collapse of the financial system in 2007 and 2008. The path of the efficiency expert is strewn with human obstacles.
Head is rightly scornful of the application of Computerized Business Systems to academic life in England, where promotion now depends on academics fulfilling Key Performance Indicators (KPIs) based on “Balanced Scorecards” of desired outputs. The value of academics’ work is now judged on publication rates, “indicators of esteem,” “impact,” and other allegedly quantitative measures. Every few years in the UK, hundreds of thousands of pieces of academic work, stored in an unused aircraft hangar, are sifted and scored by panels of “experts.” The flow of government funds to academic departments depends on their degree of success in meeting the prescribed KPIs.
China provides a hospitable venue for Computerized Business Systems and Corporate Panoptics because these are simply corporate variants of centralized control that mimic the Communist Party’s own system. Outsourcing production to China is logical for Western corporations scrambling to reduce costs, because a Chinese workforce can be subjected to pressures that, Head writes, “have no parallel at Amazon or indeed in the United States and are bound up with China’s status as a still overwhelmingly agrarian economy where the citizen is subject to the arbitrary powers of the state.”
Finally, Head shows how a “concierge” economy for the very rich has grown up side by side with the defective service economy for everyone else. In trying to escape mass-produced services, many wealthy people have turned to concierge-like doctors, bankers, etc., with whom they can have more personal relationships. “In the concierge economy…information systems are used to supplement rather than replace the skills of employees. There are no digital scripts at the Goldman Sachs private bank.”
Head offers a powerful indictment of contemporary Anglo-American capitalism, in its twin aspects of degrading workers’ skills and reducing their relative rewards. Two hundred years ago the Luddites understood perfectly well that if you deprive people of the chance to exercise their skills you reduce their earning power.
What discredited Luddite-type arguments is that the Industrial Revolution in fact revolutionized living standards. First, machines cheapened the cost of producing and transporting necessaries. After World War II, strong labor unions pushed up wages in line with productivity. The structure of twentieth-century corporations, with their multilayered systems of control, expanded the size and rewards of the middle class. Also, increased access to education had a big part in offsetting the de-skilling effects of the division of labor.
So it is not technology as such, Head argues, that springs the Orwellian trap, but the “business culture.” To show that such culture can be both humane and efficient he cites codetermination and labor–management partnerships in Germany, employee participation in software design in Norway and Denmark, the Mondragon cooperatives in the Basque region of Spain, the John Lewis Partnership in the UK, and Lincoln Electric in the US. In each case business practices enhance, not degrade, workers’ skills and pay. But from the perspective of global capitalism, generally, such practices have become rarities. So what is to be done? Politics, Head insists, needs to create a “dominant coalition” of economic losers to overcome the “new authoritarianism of the digital age.”
Head’s indictments are impressive, but at heart his book is a lament for a vanished world, that of the 1950s and 1960s, when gentlemen were still gentlemen, when digital controls were still in their infancy, when manufacturing was the main occupation of Western workforces, when services were still personalized, and when academics were paid to think, and not to produce useless papers to meet Key Performance Indicators.
Of course, in Britain (at least) the gentlemen were grossly incompetent, the cars often broke down, the services were often grumpy (expressing something of the British attitude to such things), shops opened late and closed early and stayed closed over the weekends, the food was mostly dreadful, consumer goods shoddy, consumer durables clunky, and the unions rampaged. Still, there was something undeniably “human” about those years that gives them an elegiac charm.
In his effort to recreate a worthwhile world of work, Head never faces the possibility that the machines may be destroying jobs permanently. If digital control systems can reduce workers to something close to human robots, why not dispense with unruly humans and just have robots—that is, complete the automation process? In 2012, Amazon bought a company called Kiva that makes robotic sorting systems for warehouses. This suggests that they are eventually hoping to automate even the badly paid, carefully controlled warehouse jobs that are still open to humans.
The monitoring system would then be reduced to having an engineer service the machine every so often. This would be the only human element remaining. It is currently cheaper for the company to employ lots of minimum-wage human workers to trudge fifteen miles a day between items in warehouses and packing stations. But at some point robots will surely win this race. There will be no pressure on managers to share with the robots the productivity gains they make possible.
The philosopher Hubert Dreyfus famously argued that artificial intelligence cannot mimic higher mental functions. No activity that requires intelligent behavior can be done by computers, he wrote, because algorithms cannot adequately structure the complex situations that are addressed by intelligent thinking. However, in most of the business activities described in this book, no intelligent behavior is required of most workers: the intelligence is provided by the managers; the workers only have to follow the rules of highly simplified situations. I see no reason in principle why the rules of behavior for such situations cannot be followed by machines. Algorithmic programming is bound to be much less successful in situations involving person-to-person transactions, but the number of these required for the efficient conduct of contemporary business—the production and consumption of goods and services by and for the masses that constitute the modern economy—may be shrinking.
Recently, Michael Scherer, a Time magazine bureau chief, received a phone call from a young lady, Samantha West, asking him if he wanted a deal on health insurance. After she responded to a number of his queries in what sounded like prerecorded fashion, he asked her point-blank whether she was a robot, to which he got the reply “I am human.” When he repeated the question, the connection was cut off. Samantha West turned out to be a system of recorded messages that were part of a computer program created by the brokers for health insurance.
The point is not that humans were not involved, but that the experts had worked out that far fewer of them needed to be involved to sell a given quantity of health insurance. Orthodox economics tells us that automating such transactions, by lowering the cost of health insurance, will enable many more policies to be sold, or release money for other kinds of spending, thus replacing the jobs lost. But orthodox economics never had to deal with competition between humans and machines.
Simon Head wants to make humans partners rather than competitors of machines, in a manner that restores the dignity of work. But he underestimates the scale of the reversal of contemporary business culture needed to retain worthwhile jobs for the majority. In his futuristic essay Economic Possibilities for Our Grandchildren (1930), John Maynard Keynes suggested an alternative future. “Technological unemployment”—the unemployment caused by machines replacing human labor—would continue to grow. The challenge was to convert it into leisure.
Less (human) work, less consumption, more leisure is the lesson for the future I draw from Head’s probing book. It is not the lesson he intends. But it seems to me more attuned to the possibilities opened up by continuing technological progress than does his nostalgia for a world of “good” jobs.