A Future Lacking Automatized Channing Tatum

Can you imagine a world where athletes are automated within the next 20 years? What about actors?

I, for one, cannot. After all, what would be the point of seeing a movie if all of its audio and video were not “real?” Furthermore, while I can potentially appreciate two gigantic robots straight out of Pacific Rim pummeling each other into tiny metal bits, how could such a feat be accomplished without some sort of human controller, and why would this diminish the ability of other sports to entertain?

Ultimately, I do not believe there is a 37.4 percent and 28.3 percent chance all actors and athletes are respectively going to be mechanized in the next 20 years. If they were, what would be the point of viewing them?

That is, in part, why findings such as 47 percent of all total U.S. employment being at risk of automation in the next 20 years are plainly incorrect.

Another glaring (and far more important) omission, as the leading scholar on automation induced unemployment, MIT economist David Autor, emphasizes is the fact many of the activities humans accomplish, including acting, are far more complex than they first appear.

Polanyi’s Paradox

One major flaw with computers is that they, much like horses, are ultimately not humans. This has several important implications:

  • Computers lack common sense.
  • Computers cannot improvise or solve for unexpected cases.
  • Computers are unable to adjust for flaws that were the errors and oversights of their programmers.
  • Computers are fundamentally incapable of thinking for themselves.

In other words, a programmer must first fully understand the process and series of steps fully required to perform any task they wish to code. Additionally, they must then be actually capable of writing the code for said task, which is often an exceptionally difficult feat for even basic functions.

All of this effectively means robots are good at accomplishing routine and orderly work, while the cost to perform such feats has plummeted.

Yale economist William Nordhaus estimates the real cost of performing computations has fallen by a factor between 1.7 trillion and 76 trillion (depending on the standard used). Moreover, that was in 2007.

As a result of this immense reduction in costs, firms have significantly greater incentive to implement rapidly cheaper computers.

What Does It All Mean?

A substitution effect occurs from this incentivization.

Firms will frequently switch to increasingly cheap computers in lieu of costly human labor. Autor states, “As the price of computing power has fallen, computers have increasingly displaced workers in accomplishing explicit, codifiable tasks— multiplication, for example.”

Data entry, bookkeeping, clerical work and other similarly middle-skilled and repetitive professions have been the primary casualties of automation and will continue to be automated to greater extents in the future.

However, it is extremely important to note this is only because all of these procedures follow clear-cut processes that are readily understood, thus possible to automate. There is a real limit to what can be automated, because programmers are unable to construct any program for a process they do not explicitly understand in its entirety.

This restraint is actually far more constricting than it first seems. As Michael Polanyi said, “We can know more than we can tell.”

When I am preparing French toast for dinner (because that’s how I roll), how do I know the correct way to examine my eggs to see whether or not they are spoiled? How do I turn my arms and move my hands to pick up and then crack an egg? How do I effortlessly observe the egg obviously comes from a chicken, while I add and then stir cinnamon and milk into the mixture? How do I know I want to add a hint of sugar because I have a sweet tooth that evening? How do I accomplish all of this while singing Lovefool by The Cardigans? How do I know the proper time to flip the toast when I have begun to fry it? How do I know the actual mechanical process involved in flipping a piece of toast? How do I know when to add just the right amount of butter? How do I write about this on my computer well after the fact? How do I properly construct a paragraph? How do I write in a manner people will find convincing?

It is easy to have instructions and guidelines for such activities, but all require a degree of improvisation, common sense, and intuition – abilities we only tacitly understand.

Robots – Too Stupid to Figure Out What a Chair is?

Autor emphasizes the example of recognizing chairs.

For a human, we can recognize that an object is a chair merely by glancing at it and having some sort of instant understanding that it has the properties of a chair. Meanwhile, a machine can only recognize a chair by comparing similar objects and recognizing physical features that are similar to a chair.

Autor writes, “Contemporary object recognition programs do not, for the most part, take this reasoning-based approach to identifying objects, likely because the task of developing and generalizing the approach to a large set of objects would be extremely challenging.”

Humans virtually without thinking deduce the purpose of objects and their properties, rather than solely determining their use based on appearances in contrast to other objects.

In other words, one of many reasons why robots will likely never be a complete substitute for people is because they are and will likely forever continue to be too dumb to figure out what a chair is.

This will be true at least until/if we reach a point of singularity, a point where robots can learn like humans and robots can self-improve and self-replicate, something few economists think we will reach.  However, that is a discussion for another time.

To be honest, I am not sure whether or not this is a chair either. (Image courtesy of Holger Ellgard via Wikimedia Commons.)

Robo Magic Mike?

It is true that many thinkers dating before John Maynard Keynes, who coined the term “technological unemployment,” have worried about the harms of technology substituting for human labor. However, these concerns are often far overstated, as Polanyi’s paradox demonstrates that there is more than meets the eye to much of what humans accomplish.

Although I am generally one who hesitates to forecast the future, I would say 20 years from now, as the 6th film in the Magic Mike series – now beloved in the same cinematic echelons as Star Wars and The Godfather – is released, and millions of Americans flock to see its…plot, there is a 0 percent chance the film is automated.

Not solely because the…plot “means less” if it is automated, but acting, like most human activities, is far more intricate than it first appears.  Capturing, communicating, and evoking emotions are effectively non-programmable, yet every human is able to do so without so much as a thought.

As machines gradually render rote jobs such as bookkeeping into obsolescence, it is important to note how much humans achieve that engineers cannot even imagine how to design.

So no, Channing Tatum should not fret over automation. Matthew Stafford, too, has no reason to fear being replaced by a robot, but neither does Joe the fry cook.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s