robots are everywhere
Like just about everyone else (except my wife, who is a former president of the local chamber of commerce in our small home town), for years I’ve gone to the ATM instead of a bank teller. I don’t photo checks into our account, however, although close to 10% of American check volume is now processed this way.
I see the car commercials where computer-controlled cutting and welding machines are the ultimate symbols of manufacturing excellence.
I saw IBM’s Watson trounce those two guys on Jeopardy.
So, yes, I know that robots are taking over some tasks previously done by humans.
jobs at risk
What I didn’t know is how many jobs are potentially at risk.
Then I read an opinion piece by Martin Wolf, the chief economist of the Financial Times. It’s titled “Enslave the Robots and Free the Poor.” Like anything Mr. Martin writes, the article is worth reading. But I mention it here because it references a paper by two professors from Oxford, Carl Frey and Michael Osborne, “The Future of Employment: How Susceptible are Jobs to Computerization.”
The answer is “very.” The paper concludes that 47%–that’s right, just about half, of the jobs now done by humans in the US are likely targets for replacement by robots.
How can this be?
Mssrs. Frey and Osborne divide work tasks into a matrix, according to whether the they require manual or cognitive skills, and whether they are repetitive or are non-repetitive, i.e., require some creativity, judgment or persuasive ability.
What we see in the ATM and the welding machines is repetitive manual tasks already being done by robots. We;re all used to that. The Frey/Osborne assertion is that while robots may increase their penetration of this segment of the matrix, computer scientists have become skillful enough in their algorithm fashioning that robots can now replace humans doing routine cognitive tasks. These include cashiers, waiters, tickettakers, manners of information kiosks, legal writers, medical diagnosers, truck drivers…
Is anyone safe?
Thank goodness, yes. On second thought, “Thank goodness” may not be appropriate.
–one set of “safe” jobs consists of service work that pays so little that savings don’t cover the cost of building the robot. Ouch.
–the other “safe” jobs re the ones that require a high degree of education, or that depend on creativity, or the ability to lead/persuade others, or the flexibility to respond effectively to novel situations.
fending off the robots
In the Frey/Osborne research, the two most effective ways to prevent your own robotization are to have a college degree or to be paid very poorly. Those lucky enough to qualify on both counts can breathe a sigh of relief.
timeframe
The Oxford paper gives no timeframe for this displacement. But even if the authors are off by a mile in their 47% and even if the process they describe takes half a century, substitution of capital for labor will continue to be a drag on job formation for a long while.
Frey and Osborne point out that ten years ago academics maintained that the safest possible job was being the driver of a motor vehicle. And then along came the Google car.
IBM is refocusing itself to emphasize development of Watson, which is already being used to help make medical diagnoses.
Ironically, the current ultra-low interest rate regime in the US lowers the cost of investment capital—and therefore also lowering the breakeven point that must be reached to make the investment in robots.
investment significance?
Mr. Wolf’s op ed imagines the possible long-term societal implications of further mass replacement of humans by robots. As an investor, my thought is that it may be wrong to look for the usual cyclical signs of vigor returning to the economy–signs that may never come. Safer to focus on secular growth ideas,