Merging With Technology

My Robot Wants Your Job—NO

My fellow bloggers, Ted Lewis and Rodrigo Nieto-Gomez, claim robot automation of knowledge work is proceeding at such as pace that many workers will be displaced from jobs, and there will be no new replacement jobs available for them. Robots will take over all kinds of work and there will be no work left for humans. They further claim the only good way to address this inevitable socially undesirable situation is “basic income guarantee” (BIG). There is a more optimistic side to this complex issue and it is not as bleak and prohibitively expensive as BIG. To show it, I will briefly examine these questions:

  • What is new about job displacement with AI technology?
  • How fast is it happening?
  • What is a good social response to the disruption?

What is New About Job Displacement?

Ours is not the first time in history when a large number of people were concerned machines would eliminate jobs and put many people out of work. It came up as “the machine question” during the early years of the industrial revolution about 200 years ago. It eventually faded as people discovered the industrial age generated more jobs than it replaced, and displaced people found new work. A 2013 study by Carl Benedikt Frey and Michael Osborne of Oxford University reflected a renewed interest: They found 47 percent of jobs in America are at risk of imminent automation by computers. The corresponding number for Britain was 35 percent and for Japan 49 percent. High- skilled and low-skilled jobs were least threatened; the ones in the middle most.

Behind this question is what economists call the fallacy of fixed work: There is a fixed amount of work needed to keep civilization going—growing crops, distributing food and other goods, and transporting people. If machines take over these jobs, unemployment will rise, and salaries for the remaining jobs will go down as more and more people compete for them.

Those who are worried about the current wave of AI technology believe the situation is different. Older technologies automated jobs in narrow sectors, particularly the most dangerous jobs. Newer AI technology, they believe, is capable of automating any job, blue collar or white. Therefore, the spreading wave of AI automation threatens all jobs and there is no way to get ahead of the wave.

History has shown that each new wave of technology destroys jobs of people who do the same work at a higher price, but also creates many new jobs around the new opportunities generated by the new technology. There has always been a net gain. When the previous machine question was being debated, it looked like machines threatened most jobs since most jobs at the time involved physical labor. Eventually what has been called “knowledge work” became more widespread and generated new jobs that could not be performed by machines. Today, knowledge work is threatened by the new class of AI machines. I find it hard to believe the same thing will not repeat. We already see AI technologies are limited to the specific domains in which they were trained, and humans teaming with machines can perform tasks that no machine can. There is a new category of work emerging, let us call it “design work,” that deals with creative ways to overcome problems caused by technologies that do not function as intended. Design work is already becoming prevalent and is driving the development of AI technologies. It is not knowledge work and is beyond the capabilities of machines.

Seeing that AI technology could theoretically take over all knowledge work, pessimists believe AI technologies will destroy jobs in all sectors, affecting everyone, and not be limited to defined sectors as in the past. On the other hand, heartened by history and the emergence of design work, optimists believe new jobs will appear faster than AI technologies destroys old ones.

The Economist asks, “Who is right? the pessimists (many of them techie types) who say this time is different from the past and machines really will take all the jobs, and the optimists (mostly economists and historians), who insist that in the end technology always creates more jobs than it destroys.”

Eric Brynjolfsson and Andrew McAfee are optimistic [1]. Although neither they, nor anyone else, can prove new job creation will surpass existing job destruction, they are confident history will be repeated. They also are confident that the changes in society will be slow enough that we can adapt gracefully. Yuval Harari, on the other hand, is pessimistic [2]. He believes the convergence of AI, bionics, and genetic engineering will lead to a world in which most humans are “useless” because new robots, some pure machine and others part human, will produce everything humans need for their subsistence and governments will have to provide subsistence support for them.

How Fast is Displacement Happening?

Another aspect of the problem is the changes in society fostered by computing technology are happening much faster than changes a century ago. It is not unusual for someone’s early-career skills to become obsolete two decades later. You can become a master at something after several decades and find that no one cares about it. This is a shock to your identity and sense of worth in the world. The ideal preparation for such a pace of change is to constantly learn new things. That has been hard to achieve because our education system provides insufficient support for retraining and education in new fields, and many employers do not support the continuing education of their people [3]. There is a strong consensus that employers need to step up and integrate continuing education with work and not allow their employees to become obsolete.

Brynolfsson believes institutional inertia and government safety nets will slow the processes of job destruction down to a manageable pace, allowing governments to take care of those displaced and in particular helping re-educate them.

What is a Good Social Response?

According to David Freedman, most of the talk about BIG is coming from techies. Free money seems like an easy solution. But many studies show that people receiving free money seldom use it to educate themselves with new skills; instead they watch TV more and hunt for jobs less. He wonders out loud if techies are feeling guilty that their technology is displacing jobs and the free-money proposals allow them to empathize with the problem without having to do much other than contribute money.

And the money needed to finance BIG truly is BIG. In the U,S. about 200 million adults ages 15-65 would qualify for the subsidies; at $18K per adult, the total bill would be $3.6 trillion—the same as the whole U.S. budget. No one knows where that money would come from. Voters are not ready for this at all: in Switzerland in 2016 they soundly defeated a BIG proposal with more than 75 percent disapproving. Freedman quotes Brynjolfsson : “There’s still plenty of unmet needs and work to do, so the right strategy for the current situation is to prepare people for those new tasks.” And “For now, we’re not rich enough to afford a basic income that will provide everyone with a decent standard of living without having to work.”

Education supplemented with government safety nets is by far the preferred approach.


AI technologies bring both risks and benefits. Much of the discussion has focused on the risks and has ignored the benefits that leave everyone better off and foster a new kind of work—design work. Three prominent examples of AI-assisted technologies that will provide a big benefit to everyone are fully automated driverless car systems, increased interactivity between people and machines, and significantly accelerating scientific and medical research. Designers are in high demand in all these areas. There is more to welcome than to fear.


[1] Brynjolfsson, Eric, and Andrew McAfee. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies.   Norton, 2014.

[2] Harari, Yuval. Sapiens: A brief history of humankind. Harper, 2015.

[3] Gordon, Edward. Future Jobs: Solving the Employment and Skills Crisis. Praeger, 2013.