A recent article by The Economist highlighted the potentially fundamental changes that may be coming to a work near you. Read the Difference Engine: Luddite legacy (1)! It may help you decide how threatened you should feel by new advances in computer AI.
Five years ago, I read a book by an economist Jermy Rifkin, titled provocatively, "The END of WORK, The Decline of the Global Labor Force and the Dawn of the Post-Market Era". Prevailing wisdom at the time (~circa 2000) was that 'creative destruction' of jobs mean workers have nothing to fear from technology. The thinking is that older jobs, like assembly line work, would be replaced by cheaper and more productive robots. But in this process of destruction, new jobs are created, such as the repair, design, maintenance of those robots. As long as workers can retrain and retool, they'd simply skip up the food chain to better paying skilled jobs.
Rifkin however, argues that: As progress and productivity marches inevitably forward, we're loosing more jobs to technology than we're replacing with jobs to service that technology.
At the time I found his statistics backed arguments vaguely persuasive, but it didn't hit home until recently...
Fast forward to the Great Recession of 2008 - 2011 (now)
It is now November 2011, the U.S. unemployment rate is still at 9%, the latest October job data shows the U.S. economical machine added a mere 80,000 new jobs. Europe is in a boatload of trouble stemming from their sovereign debt crisis, but also stemming fundamentally from the lack of productivity per capita in a subset of Euro-zone countries. China can't export to either of their biggest traditional trading partners, and has embarked on a massively expensive infrastructure investments whose consequences we have yet to witness.
With so much economical uncertainty, its no wonder workers of all collar colors are feeling the strain. So why do corporate financials showed record-breaking profits and productivity gains? My back-of-the-napkin calculations using the U.S. Bureau of Labor Statistics (2) productivity index suggest that companies have increased their productivity almost 10% since 2008. And since Americans aren't OT'ing 10% more since 2008, where are the gains coming from?
New productivity gains are from machines (not people)!
The New York Times (3) pointed out that after the Great Recession, companies are happily shelling out big bucks on hardware and software (steep exponential green line), but not on new (or existing) workers (flat blue line). There are many reasons why workers are more expensive than machines (the cost of healthcare and pension benefits for example). For this reason companies will keep investing in machines until they run into problems only people can solve, and only then will they hire more workers.
Is machine learning making people obsolete?
Language:
In the objective of providing services to people, the current bottleneck to have machines service people directly is the natural language barrier. Simply put, machines can't talk intelligently to people yet, nor understand people well enough when they talk back. But great strides have been made in recent years to address this issue.
Apple's Siri is already providing a spirited (albeit buggy) attempt to create a cheap secretary for the masses. You can be sure as Siri collect the oodles of data from current users, the Siri of the future will get spookly better until you can't tell the difference from a real secretary vs your phone. And where Apple goes, you can bet your lunch that Google and Facebook are also racing to use their massive stores of user data to create their own attempts to break the language barrier. Once machines can accurately understand the natural language, millions of service jobs relying on language expertise will become redundant overnight.
Medicine:
Recently my wife had her hand CT-scanned to see how well a recent medical procedure is mending. The radiologist was nice enough to let me observe the whole procedure. From my (non-medical) perspective, he simply pressed a button, and the CT machine automatically scanned up and down her arm. When the results came back, the computer highlighted the her cast and allowed the radiologist to exclude that part for the final report. When I asked him about the condition of the hand, he replied with a standard 'talk to your doctor' routine (although he was very professional about it).
This was a seasoned radiologist with over twenty years of field experience and a six-figure salary, but the only thing I saw was his photoshop skills as he methodically deleted the cast from the report. I left his office wondering if radiologists would become redundant once the CT system can do their own photoshopping. The Canadian Radiology organization and doctors would probably object, so those jobs are safe for now, but it makes you wonder what jobs in the medical field can be eliminated with the proper incentive?
Bottom line:
These anecdotal evidences suggests that in the pursuit of the bottom line, companies will not hesitate to outsource, or outright eliminate jobs using cheaper technology. Evidence also suggest that machines are becoming more capable of doing jobs currently done by people, from fields as diverse as language and medicine. While the geek in me celebrates "progress" and our increasing standards of living, the Luddite in me hopes the jobs I'm good at aren't machine replaceable.
There's a story floating around about the time Ford CEO, Henry Ford II showed Walter Reuter, leader of the United Automobile Workers a newly automated car plant and said "Walter, how are you going to get those robots to pay your union dues?” Without skipping a beat, Reuther replied, “Henry, how are you going to get them to buy your cars?”
To put it another way, in our drive to create life-improving products and services with technology, are we impoverishing the very customers we're hoping to attract? It's definitely a dilemma of our generation.
A prototype office robot coming to a Japanese office near you. "Yes yes, laugh at me now, but I'll eventually take over your job! Mu-Ha-Ha" Says the Evil Robot Office worker as he does the evil robo-dance.
References:
(1) Difference Engine: Luddite legacy - The Economist
(2) Major Sector Productivity and Costs Index -U.S. Bureau of Labor Statistics
(3) Man vs. Machine - New York Times
