Do the Robot

George Condo, Jean Louis Mind (2005)

Good news for the labor theory of value, from a recent article from the Harvard Business Review: "Automation Won't Replace People as Your Competitive Advantage." That's meant to be reassuring. Even if we can't imagine any jobs so specialized or service-intensive that a robot can't steal them, we can rest assured that capital won't always find it profitable to go with the machines. Humans, it turns out, have creative and affective faculties beyond strict functionality — the article's authors, drawing on a new management tract called Humans Are Underrated, identify "empathy" and "storytelling" — and companies will need to find a way to take advantage of these to beat out competitors. At the end of the day, you can't really exploit a robot the way you can a person.

As the authors of the article explain, "You will always need good people. And you need a system that engages them and allows what is unique and valuable about individual people to be leveraged." Being a "good person" means exceeding the job description in ways that profit the company. That is the essence of "individuality": It is not a product of our snowflake souls but a production input shaped by the exigencies of the capitalist production process. "Uniqueness" is simply what can't be otherwise automated but remains exploitable.

The corollary to that is that we work to produce "unique" aspects of our personality to become useful to capital. (This is the gist of the Marxist idea of the "real subsumption" of identity. Capitalism's mode of production fully drives the formation of personality rather than adapting its processes to some inflexible pre-existing identity-formation process.) The culture-wide celebration of authenticity is not a revolt against corporate values but an expression of them. Authenticity is precisely the opposite of the image of a disinterested, spontaneous self that the word sometimes conjures. If you are not "authentic" enough to be exploitable in some way — if your personality can't be "leveraged" — then authenticity is not really available to you. You can't afford to be yourself.

The unpaid work of striving for "authenticity" is a matter of formatting oneself into an exploitable human resource, into a kind of rival good that one company can make use of to the detriment of another. One of the problems for capital with automated processes is that they distill the hidden abode of production into a kind of intellectual property, which makes it eminently copiable. Thus, the advantages of such processes readily spread to competitors: They just copy your machine, install your software.

Thus automation leads to a new urgency for employers to exploit the unique qualities of their employees. The HBR article urges managers to recognize this:

In our highly competitive economy, managers may be too easily seduced by the apparent advantages of automation. In relentless pursuit of lower costs and greater throughput, they might miss the fact that advantages in storytelling, judgment, and other human strengths are much harder for competitors to replicate.

It's nice to see that storytelling and human judgment being championed here, but we shouldn't kid ourselves that the management class would fight for those qualities if they weren't profitable and couldn't be put to competitive advantage. The minute a machine could tell a story that reliably sells, the management gurus would be telling us to "Get used to it, tomorrow's Shakespeares will all be robots," as robot stories flooded the market and destroyed the ability for  a human to write for a living, and then the search would be on for some other discrete "human strength" to reify and exploit. What constitutes a human strength is highly malleable and contingent on machine capabilities; it is driven by and measured in terms of what machines can or cannot simulate. It is not a matter of some sort of essential species being.

Regardless of what the human strengths happen to be at any given time, companies will do their utmost to maximize the value of those "uniquely human qualities," touting them as marks of authenticity and management's devotion to both consumers and workers. Contrived cheerfulness, because it is currently hard to automate or convey with machines, is heralded as "authentic" or "caring." The authors cite the example of Southwest Airlines, because the company attempts to differentiate itself through warm service, "hiring for attitude and training for skill." The authors of the HBR article note that the company "doesn’t lack for press about its positive organizational culture and cheerful customer-facing employees," and a Google search certainly confirms this. But it would be a mistake to see the qualities that Southwest deliberately tests and hires for (positive attitude, creativity, sense of humor, other-orientedness) as "authentic" human qualities that they merely screen for and nourish. Rather, their hiring process produces those qualities as valuable, establishes a material process by which the fabrication of those qualities becomes a useful job for an employee to assign themselves. (Ironically enough, that hiring process increasingly makes use of algorithmic sorting and Big Data; machines are used to help select the most valuable unique and unmachine-like humans.)

In other words, Southwest sets up a system where it behooves workers to pretend they care for customers in particular ways, to work hard at the simulation of care, which basically precludes the possibility of the care being "genuine," if that even means anything. Arlie Russell Hochschild's The Managed Heart details the nature of this "emotional labor," which she argues impinges on a worker's ability to experience emotions that are their own and not instigated by the job. When workers try to force "authenticity" for work, it precludes their ability to feel authentic about themselves in any context.

The behavior that seems "genuine" to customers is not fixed — the folksy preflight routine in the aisle can quickly come to seem stale and corny. What makes a person seem genuine in a commercial context has nothing to do with the actual disposition of the people involved in an exchange but with the expectations established by other commercial interactions. This follows from the logic of people's "uniqueness" being seen solely as a competitive advantage, something that consumers covet and can detect. "Genuineness" is a real thing only to the extent that it is differential from other companies' phoniness. There are no "genuine" interactions, only ones that seem so and keep the customer satisfied.  Profit determines whether behavior is genuine: If it was profitable, it was genuine; if not, it was phony.

Ways of seeming genuine get depleted through overuse, and new ways to convey sincerity must be invented. Employees must not lapse into a routine and "go into robot," like the disgruntled flight attendants in Hochschild's research. But at the same time, the "sincere" emotions must be accessible on demand, as part of an industrial process. Southwest screens not for people who can't go into robot, but for people who go into robot in less detectible, less familiar ways. In a way, it would suit companies if people would give up on the attempts at faking genuine feeling and believe that they can simply and unreflectively "be themselves": This would preserve the integrity of companies' hiring and screening systems and assure that they won't be gamed. Then companies could hire people who don't have to fake what passes for genuine at that particular juncture. They could always replace these savants when the way they are naturally begins to seem contrived to others. People who make no effort to be themselves are perfect for validating the "objectivity" of personality tests. Advice to "just be yourself" translates as "Don't you be anything, let us tell you who you are!"

But once beyond the hiring phase, companies want workers who are capable of working hard at feeling, even if it is simply to mask it. Emotional labor only works as competitive advantage when it can be compelled by bosses yet remain convincing, improvisational enough to fool customers. Service businesses need to avoid making emotional labor into what the HBR authors call a "system that compels people to perform standardized acts in the same way." So bosses have to insist on an "authenticity" from workers that takes their alienation up an exponential level. Companies impose a system in which employees are held responsible for making the systematicity invisible. Workers are compelled to find ways to be creative about performing routinized tasks and seem to be organically "doing what they love." (Pret à Manger's experiments in forced spontaneity are an interesting example of all this.)

Miya Tokumitsu had a good critique of the Do What You Love ideology in Jacobin, in which she argues that "do what you love" means turn your passion into human capital — the real subsumption of identity in another guise. She writes,

 According to this way of thinking, labor is not something one does for compensation, but an act of self-love. If profit doesn’t happen to follow, it is because the worker’s passion and determination were insufficient. Its real achievement is making workers believe their labor serves the self and not the marketplace.

Having a "real" passion for your job is the extension of exhibiting "genuine" feeling in the workplace, but instead of serving a customer, it serves a boss or client. Again the metric that establishes the reality of feeling is ex post profit. If no one wants your passionate work, it's not really passionate and you are self-deluded.

Tokumitsu argues that genuinely lovable work is a privilege that comes at the expense of lots of unlovable work being done by others:

Work becomes divided into two opposing classes: that which is lovable (creative, intellectual, socially prestigious) and that which is not (repetitive, unintellectual, undistinguished).

As a result, Tokumitsu argues, unlovable work becomes "dangerously invisible" to those whom it permits to do what they love. And in the meantime, those who love what they do work harder for less or no pay.

But the logic that sees competitive advantage in the "human touch" means that all work must be lovable and be performed as such for customers (and the managers who are supposed to be their proxy). Unlovable work isn't made invisible but is made to seem visibly, irrepressibly loved. After all, what keeps a crappy job from being automated, from this perspective, is the joy in it that a worker can manifest and woo customers with. What prevents a job from being automated is not necessarily its complexity, as Peter Frase explains in this post (and elsewhere):

From the perspective of the boss, replacing a worker with a machine will be more appealing to the degree that the machine is:

  • Cheaper than the human worker
  • More convenient and easier to control than the human worker

If workers demand more wages, machines become more attractive to bosses. Likewise with "meaningful work": If workers demand more meaningful, lovable work, then they become less "convenient" to bosses. But workers whose value rests in how much they show they love their job are quite easy to control. Servility is built into the practice. Frase writes that "the truly dystopian prospect is that the worker herself is treated as if she were a machine rather than being replaced by one." Even more dystopian is the prospect of being treated like a de facto machine while being expected to express boundless "human" joy about it.

The threat of automation, then, can be used to extract more emotional labor and more competitive advantage from humans. After all, one of the few things a robot can't supply is enthusiasm.


This paper by economist David Deming (via) makes the case that because of automation, “the labor market increasingly rewards social skills.” Translated out of economist-ese and into a language critical of capitalism rather than complicit with it, that means the point of exploitation is shifting, moving deeper into the human personality and turning more of life’s nonwork and leisure into exploitable labor. Social interaction is tamed and subsumed, so that it can’t be oppositional to capital but is a “human resource” within it.

Deming notes that “high-skilled, difficult-to-automate jobs increasingly require social skills,” by which he means jobs that require “tacit knowledge” and the ability to read other people’s behavior and put yourself in their shoes. Jobs, that is, increasingly involves empathy as a discrete form of labor, empathy as competitive advantage, empathy as something bosses can compel and manage.

This will have the effect of codifying the empathy process, making it ultimately amenable to automation itself. The rising value of social skills increases the capitalist desire to automate it, save on labor costs, and then push the point of exploitation to some new and deeper level of human behavior.

The commercialization of empathy, the reification of it as a skill, also gives incentive to workers who must exhibit it on demand to optimize the effort it takes: This entails becoming emotionally simplified and demanding emotional simplification in others. It means contriving situations, work processes, in which “the other person’s point of view” is constrained to only a few conceivable ”rational” options. It means reducing empathy to routines. It means self-robotification among workers in hopes of streamlining what it takes to see from the other person’s point of view. This does the work of translating “social intelligence” into something ultimately code-able.

The model of workplace skills and team building that Deming creates for the paper actually serves as a kind of blueprint for how this could proceed: “This paper should be viewed as an attempt to extend and formalize the definition of one particular dimension of “soft” skills — the ability to work with others,” he writes. These soft skills don’t resist automation so much as exist higher up the automation chain. But the work of automating social skills isn’t a matter of more sophisticated computer programming; it’s a matter of reducing the scope of typical human behavior — and the range of empathy that people are capable of and expect — to something that can be programmed.

There's nothing inherent in emotionality that resists behavioralist reduction, that can't be programmed. It requires active belief, active resistance, to maintain that emotion can't be automated or outsourced or circulated like currency. But for that resistance not to be co-opted, not to attenuate some new form of human labor to exploit, it has to be conceived as making emotions useless. Resisting the automation of emotion means resisting emotion's recruitment as a skill.