Run for Your Lives, the Robots are Coming

​"Watson for President," a friend of mine joked, referring to IBM's prominently advertised question-answering computing system.  Lately, commercials for the system have run in what feels like a loop, a human irritant. 
 
Initially, I had expressed fear of an eventual artificial intelligence takeover, especially of the U.S. job market – and, perhaps, the human race itself.  This quip was my friend’s response, and he dove deeper into the pit of his desires for an A.I. takeover.  He was getting dark with the fantasy, and the darker he went the more he seemed to enjoy it.  I can’t say I didn’t enjoy it myself, especially in the way the others at the table had started to look at him.
 
“Just kidding,” he finally surrendered, before cackling, throwing his weight back in his chair and proudly crossing his arms.  I wasn’t sure if he was. 
 
A simple google search on the subject brings up a long list of fearful articles with titles like “Robots Stealing Jobs” or “U.S. Job Market in Jeopardy of Robot Takeover” or “Run for Your Lives, the Robots are Coming”.  Some of them bring up valid, statistically backed points.  Others don’t, too often focusing on farfetched sci-fi movie plots.  I’m not one to talk, for I, too, have brought up similar plotlines in casual conversations.  Dinner parties, usually where I wasn’t exactly invited, and was more often than not the loudest there. 
 
“Watson for President,” my friend backtracked, uncrossing his arms and retracting his momentary dissension. It was to the dismay of the others, and at the table, to my joy. My fuckin joy, as I'm sure they heard it.    
 
In an analysis of the current U.S. labor market one finds that the most common job held across the United States is truck, delivery, and tractor driver, with some 1.7 million people employed as such.  Here’s a 2014 map of the most common job in every state courtesy of NPR.
 
Regardless, when analyzing the effects that automation will have on the U.S. labor market, the specific job is unimportant.  Rather, determining how routine the job is proves more vital a factor in representing whether or not it will be replaced.  Truck driving is largely a routine job, and already, Uber, in collaboration with Otto, a San Francisco self-driving tech startup, powered an 18-wheeler filled with beer to Colorado Springs late last year.  The largest autonomous beer run ever completed, and first ever autonomous truck delivery.  Beer aside, the technology to replace truck drivers has arrived and is steadily advancing. 
 
Consider for a moment, if tomorrow all truck driving jobs were lost to automation.  That would leave over one million people without employment, and likely very angry.  What would happen next?  Let your imagination run wild, I know mine is.  Would they take to the streets?  Would they strike? Revolt?  Those are serious numbers spread out across the entire country, and with any sort of organization they very well could. 
 
There’s a sick part of me that truly hopes that I’m around when something like that happens.  I recognize the heartlessness in that, but this sort of desire is present in all humans, I imagine.  It’s the burning car on the side of the road that you can’t stop watching.  It’s the urge to reach out and touch a corpse at a funeral.  It’s Jack Nicholson axing through a door, screaming “Here’s Johnny!”  It’s referred to by psychologists and sociologists as morbid curiosity.  It’s the same curiosity that draws us to worst case scenarios, horror films, and fantasies of apocalyptic revolts. 
 
“Watson for President,” my morbid friend repeats over and over as the other around the table plead with him to stop his fantastical wandering down his path of worship to an artificial intelligence overlord.  He will not stop. 
 
Many evolutionary psychologists relate morbid curiosity to our genetic disposition for survival.  If we are able to safely watch someone else’s suffering, we are more likely to know how to deal with a similar situation ourselves, if ever presented.   Some psychologists pose that our morbid curiosity comes from a desire to feel empathy for other humans.  To connect.  
 
As I contemplate my own morbid curiosity, I’m struggling to understand which category I fall into.  To me, I look to it as an issue of inevitability.  If it’s going to happen, I might as well be there for it.   
 
What does morbid curiosity have to do with automation?  Right now, directly, not much.  But down the line, it could have everything to do with it. 
 
It has to do with the millions of truckers, postal workers, tellers, cooks, secretaries, sanitation workers, janitors, construction workers, and any other routine job that will eventually be automated.  That kind of displacement can have radical effects on society.  And the technology to get us to that point is rapidly developing. 
 
For example, Google’s driver-less vehicles covered more than 600,000 miles last year. According to economist Andrew McAfee, “We are facing a time when machines will replace people for most of the jobs in the current economy, and I believe it will come not in the crazy distant future.”
 
Is that kind of displacement possible?  Perhaps. 
 
“Google is worth $370 billion but has only about 55,000 employees – less than a tenth the size of AT&T’s workforce in its heyday [in the 1960s],” wrote senior editor of The Atlantic, Derek Thompson. 
 
There has been little dialogue discussing the implementation, workforce impact, and policy reform required to ensure a smooth transition to automation.  The Obama administration did put together a report examining certain possibilities and solutions, but it was more a surface-level examination than anything.   
 
It is very likely a thick mass of government regulations and legislation will follow automation like the residue left behind by a slow-moving slug. 
 
However, turning our focus back to driver-less vehicles there is a compelling case to be made for automation.  To start, over 40,000 people die each year in motor accidents, the majority of which are caused by human error.  Human error. 
 
Dick Schmidt, a UCLA professor and Porsche Racer, and perhaps the foremost authority on something called impulse variability, which essentially refers to the variability in movements of a person’s limbs when faced with an event or situation that sparks an impulse reaction, has dedicated much of his research to explaining this human error. 
 
In 2010, when all those Toyotas were recalled amongst the public hysteria surrounding the alleged accelerator throttle defaults reported, it was immediately concluded that the cars were to blame.  Specifically, many media outlets and even Attorney General at the time, Eric Holder, leapt to say to that Toyota had acted irresponsibly, putting profits above the safety of the consumers.  Experts in psychology like Dick Schmidt, never once felt the vehicles were to blame.  He wrote a great op-ed about it in 2010.  You can read that here. 
 
Essentially, what he determines is that in every case, in conjunction with the data collected from the black boxes in the vehicles, none of it coincided with the events reported.  These drivers simply thought they were hitting the brake, when they were actually hitting the accelerator. 
 
Schmidt explains they had not confused the two pedals, meaning that if you were to ask them which was which – they would know that the throttle was on the right, the brake on the left.  What he means is that there was impulse variability, in which the driver’s intention and action became jumbled, resulting in the mistake. 
 
“First, in these situations, the driver does not really confuse the accelerator and the brake. Rather, the limbs do not do exactly what the brain tells them to,” wrote Schmidt. 
 
This is interesting because the overwhelming conclusion (Toyota paid billions in settlement cases) was that the vehicles were to blame.  No blame was ever rendered on the drivers.  Human error thrown out the window willy-nilly.  It is this dissonance that leads me to believe that automation may be further away than we think.  Today, more than ever before, there are systems and programs and technologies in our vehicles that are designed to keep us safer, yet the number of motor vehicle deaths each year continues to rise.  Of course, there are more people in the United States than ever before, but shouldn’t these systems and technologies make some sort of mark? 
 
Give me a robot chauffeur and I’m happy as a clam.  On long road trips, I’ve been known to sleep through entire states.  See ya later, Iowa.
 
Moving away from cars, consider the long reach with which automation would swing.  Isolating how routine any one occupation is, it becomes clear that automation will not only affect blue-collar jobs, but white-collar ones, as well. 
 
Take a field like radiology, for example.  A large and important part of a radiologist’s duties are, actually, very routine in that he or she is expected to repeatedly examine CT scans.
 
Dr. Igor Barani, an oncologist and CEO of a deep learning company called Enlitics, pit his system against three expert human radiologists in what was a John Henry versus steam engine type contest.  In examining CT scans of a patient’s lungs, Barani found that the Enlitics system performed 50% more effective at classifying malignant tumors, and missed cancerous cells 0% of the time, compared to a 7% miss rate by the human experts. 
 
Add doctors to that list, I guess.  See ya later, Doc.
 
“Watson for President,” declared my friend as he finished the beer in his glass and threw it against the wall, shattering it into pieces.  Promptly, a robot security guard rolled up, tased and hogtied him, before escorting him down to the robo-precint on 43rd. 
 
He’s out on bail now for a million bitcoins, and, remarkably, he’s stilling chanting, “Watson for President.”
 
I admire his resilience, although I think it’s misdirected.  

For now, here's a list of jobs most likely to be replaced by robots.