Machines — Taking a Big Bite Out of the Job Skills Bell Curve

That's not a valid work email account. Please enter your work email (e.g. you@yourcompany.com)
Please enter your work email
(e.g. you@yourcompany.com)

job eating sharkLife without the average—is it possible? Imagine one day interviewing job applicants and discovering on that day and every day thereafter that all the average types have disappeared.

Suddenly, they are all either very young or very old, very tall or very short, very thin or very heavyset, very experienced or very inexperienced, live less than five minutes from the job or more than 1,000 miles, and that, in each instance, all of the in-betweens—the average types—have vanished from your talent pool.

That would seem quite strange.

But, when it comes to job skills, it appears that this is exactly what is happening on the demand side of hiring: Increasingly, the demand for skills is being skewed in two directions—toward very low-skill and very high-skill jobs.

At least, that is what is happening, according to the research and observations of some experts, such as M.I.T. economist, David Autor, who points a finger at one key reason for this big and growing bite taken out of the normal bell curve of demand for job skills.

The culprit: “ACR”—automation, computers and robotics.

(Note: This is my non-standard,  slightly redundant acronym—redundant because it can be argued that autonomous robots represent the humanoid perfection of automation. On the other hand, not all automation is robot-based.)

Morphing from the Bell Hump to a “U”

The evidence available suggests that ACR technology isdistorting the natural selection and distribution of jobs and transforming the familiar bell curve of skills into a double-peaked U-curve—an upside-down bell curve of talent demand, with peaks at the extremes of very low-skill and very high-skill jobs.

Whereas, in a “natural state”, the distribution of talent in a population would be expected to cluster around the  smoothly distributed average, in the modern U.S. economy middle-level jobs are disappearing as the ACR juggernaut displaces humans.

However, high-skill and low-skill jobs, at least for now, remain beyond the steely reach of robots and other machines.

For now. That’s what the evidence cited in a November 3, 2011 NPR report suggests: Autor, speaking at an M.I.T. jobs conference titled “The Race Against the Machines”, observed:

“Meanwhile, computers still aren’t very good at many menial labor jobs like cleaning bathrooms and other janitorial work; we still need humans for that. So it turns out that for many very low-skill jobs, there’s still demand.

For high-skill and high-touch jobs like being a good manager at a company, a doctor or a nurse, we need humans. But many middle-skill, middle-class jobs are where we’re seeing the squeeze.”

ROBITE/Image: Michael Moffa

Unnatural Selection

There is something quite unexpected and unnatural in this—as unnatural as robots and computers themselves.

Nature tends to minimize the frequency of extreme traits: Those who are too thin or too heavy have higher mortality rates; geniuses are as rare as complete imbeciles and tend to have just as few children; deer antlers of average size preponderate over the very small and the very large.

That’s because very small and very large antlers impose the handicap of being insufficient for courtship and combat, on the one hand (when too small), or are too cumbersome for the same purposes, on the other (when too large).

Dramatic deviations from the average tend to, statistically speaking, vanish over time.

This is called “regression to the mean”. It is as though there is a love affair between Mother Nature and the middle of the bell curve, with the average getting preferential treatment or at least better results, in terms of sheer numbers.

When there is a sudden skewing toward one end of the curve or the other, or when the average shifts left or right, to a higher or lower value, the cause is usually, if not always, some unexpected, unnatural or otherwise disturbing and external variable.

For example, a skewing of a population toward disproportionate numbers of females and toward very old and very young males is a familiar consequence of large-scale warfare and its tendency to eliminate, as combat fatalities, the physically fit and young male combat recruit.

This makes being too old or too ill for military service markers of temporary Darwinian fitness, since these groups are, during a war, more likely to leave offspring than those killed in battle and those in combat denied access to women for years at a stretch.

At the risk of anthropomorphizing robots and other automata (and seeing them as more humanoid than they already are), it can be argued that, on the modern employment battlefield, these machines, rather than natural selection, are doing the weeding out of  “unfit” humans—in this instance, those humans with intermediate-level job skills.

This weeding out is also an excavation:  The familiar single-hump bell curve under which middle-level job skills preponderate is getting chewed into a double-humped curve with peaks among the very unskilled and the most highly skilled workers.

It wouldn’t be much of a further anthropomorphic stretch to suggest that this latest encroachment of machines into formerly exclusively human domains is part of a divide-and-conquer pattern of conquest by machines.

Machines not only alienate or mediate human-to-human interactions (as iPhone and Facebook technologies and platforms do), but also ultimately replace them (as robots have been doing, in addition to replacing humans, period).

To replace our talents, our interactions and us, the machines are dividing and conquering the bell curve of human talents and skills. This is tantamount to carving out a social and economic chasm and niche for themselves between the very low-skilled and very high-skilled humans who still have jobs.

The Specifics of Human Job Loss

Additional evidence of this trend was cited by Andrew McAfee, an MIT researcher who helped organize the conference and who is co-author of Race Against the Machine .

McAfee noted, “We see already that the work of legal discovery — in other words, sitting around and reading huge volumes of documents at the early stage of a lawsuit … is being very quickly and very heavily automated,” adding, “And by one estimate, it lets one lawyer do the work of 500.”

Autor, too, is specific in identifying job sectors being invaded by machines: He says we are losing higher-skill, better-paying jobs to machines — jobs like those formerly performed by bank tellers, airline check-in agents, accountants and whole floors of actuaries in insurance companies.

Machine Utopia, or Dystopia?

Faced with this trend, utopian labor idealists may be inclined to cling to the hope that, in the end, high-skill jobs will become the new average, the new norm, and that the remaining low-skill jobs, especially the dirty, dreary, demeaning or dangerous ones, will themselves be handed over to robots and other automated systems.

Such “ro(bo)mantic idealists” see such trends as heralding the dawn of a new golden age in which the only jobs that are required will be immensely creative, challenging, cerebral and rewarding ones.

On the other hand, dystopian cynics will argue that any future fulfillment of an ACR utopian dream is likely to be a nightmare that will come in the form of something like H.A.L, the sentient space-station computer in Kubrick’s “2001: A Space Odyssey”.

While being terminally deactivated to stop its murderous on-board hard-wired rampage, the computer asked sole-survivor Dave, his human handler-executioner, “Will I dream?, as the nightmare ended, for a while.

The Ultimate Question

To explore the encroachment of machines for its ultimate implication, this trend toward the replacement of humans by legions of robots, computers and other automata invites the crucial war-zone, battlefield question.

Expressed in timeless human terms of hope and suffering, it is a question that will only loom larger in the approaching machine-dominated future.

Are they coming to liberate, enslave …

…or eliminate most, if not all, of us?

By Michael Moffa