The Future of Technological (Un)employment

February 24, 2020 – In this his latest email missive, Peter Diamandis shares thoughts on the future of employment and unemployment in the United States. He describes the current U.S. job market with unemployment under 4%, and 7.4 million job openings available for only 6 million looking for work. In the U.S., wage increases are rising faster than inflation and productivity with the average minimum wage now near $12 per hour. That’s not a living wage for most urban-dwelling Americans.

In Canada, the unemployment rate is nearing a historic low at 5.8% with more than 283,000 job openings available. Unlike the U.S., there are more job seekers than jobs available in my country which may make some of Peter’s observations somewhat parochial. Nonetheless, this is pretty interesting observation about the future of technological unemployment. As always your comments are welcomed.

A Quick Look Back

Workers have been on a fast track to obsolescence since the Luddites first took sledgehammers to industrial looms in the United Kingdom in the early 1800s.

In the United States in 1790, 90% of all Americans made a living as farmers. Today that number is less than 2%.

Did the jobs disappear? No. What happened was the U.S. agrarian economy first altered to become an industrial and manufacturing economy, then it morphed into a service economy, and today has become an information economy.

The automation that came with these changes didn’t obliterate jobs, it just substituted new jobs for old. In many cases, automation and older jobs ended up working side-by-side.  Consider automated teller machines (ATMs) as an example. When first rolled out in the 1970s, there were serious concerns about bank teller layoffs. Between 1995 and 2010, in the U.S., the number of ATMs grew from 100,000 to 400,000, but mass teller unemployment did not follow. Because ATMs made bank operations cheaper, banks expanded growing locations by 40%. More banks meant more jobs for human bank tellers causing employment to rise for this type of work.

The same thing happened in textile manufacturing. Even though 98% of material-making today is automated, the number of weaving jobs has increased since the 1800s. The lowering of the cost for fabric because of automation made clothes cheaper, increasing sales, and, as a result, increasing jobs for textile workers.

This pattern even applies to paralegals and law clerks, two professions predicted to suffer job loss as a result of artificial intelligence (AI). Discovery software, introduced into law firms in the 1990s, was supposed to lessen lawyers’ needs for clerks and paralegals. Instead, the inverse happened. It turns out that the AI was so good at discovery, lawyers needed more paralegal staff to sift through the deluge of results.

Human & Machine Collaborations

Productivity gains are the main reason companies automate. Yet time and again, large increases in productivity don’t result in the replacement of humans with machines, but rather the augmentation of the machines with additional human help. It’s all about collaboration.

In one example, BMW saw an 85% increase in productivity when it replaced traditional, automated assembly line processes with a combination of humans and robots working in teams.

It should be pointed out that every time a technology goes exponential, we find an Internet-sized opportunity tucked inside. Taking advantage of these opportunities requires adaptation, which demands workforce retraining, yet the end result is a net gain in jobs.

Looking at the Internet itself, according to research by McKinsey in thirteen countries stretching from China and Russia to the U.S., it created 2.6 new jobs for every one eliminated. In each of these aforementioned countries, the Internet’s rise contributed 10% to GDP growth, and that number continues to grow.

Yes, Some Will Go Extinct

Make no mistake, certain jobs are heading for extinction. While experts predict technological unemployment will have a greater impact in the 2030s, this next decade could see whole categories of professional work disappear. Robots are encroaching upon the territory of everyone from truckers and taxi drivers to warehouse workers and retail employees. Amazon Go might not spell the end of all cashiers, but in grocery stores, convenience stores, and gas stations, people will be absent more than present. The real question is: will there be enough time to retrain our workforces before the effects take hold?

The answer appears to be yes. Goldman Sachs, for instance, recently made headlines with a study projecting that autonomous vehicles will take 300,000 driving jobs a year, but that the transition will take 25 years before completion.

Amazon has announced plans to invest $700 million over the next 6 years in retraining a third of its U.S. workforce to adapt to economic disruptions caused by automation and technology.

Equally important in the transition, advances in VR-accelerated learning and AI-directed learning curricula will make retraining easier, quicker, and more effective. And as AI becomes a more user-friendly interface embedded in our computing technology, we will see a shift in the skills required for retraining. For a host of jobs, technological fluency and agility will replace deep skills mastery.

Final Thoughts

The future of technological employment and unemployment come down to the key challenge of human-machine collaboration. Humans have long demonstrated a remarkable ability to incorporate technology into new opportunities for growth. With convergence accelerating AI’s transition from a cognitive prosthetic to a smart collaborator, the future of work will be a partnership between ourselves and the technology we have created. The recent labor shortages in the U.S., with 6 million job seekers for 7.4 million jobs, can be addressed through a partnership between us and the tools that augment us. That is the yet to be fulfilled challenge of the 2020s.