AI Will Not Eat All the Jobs

In 2014, there was much fear and excitement about the technologies of the time, just like the current hype around ChatGPT, and back then, like now, everyone was talking about how technology was soon going to eliminate a large chunk of the jobs in the economy. I wanted to better understand the issues, so I did a deep dive on the topic. After reading and thinking through things, I posted a series of blogs on my conclusions. I just re-read my analysis, and I think all the points I made back then still hold true today. In particular, I concluded that, perhaps after periods of adjustment, that new technologies, would continue to create many more jobs than they destroyed.

In the first part of my earlier analysis, Two Schools of Thought, I established that people opining on the subject generally held one of two perspectives regarding the expected impact of new technologies on future labor markets. (At the time, there was a trend in which middle class jobs were slowly being eliminated from the economy, and that same trend has generally continued on through to the present.) The two perspectives were:

• The “This Time Is No Different” view suggested that, based on what had always happened in the past, we could expect new technologies in the future to eventually create more jobs than they eliminated.

• The “This Time Is Different” view suggested that further improvements in technology would further exacerbate the trend toward a bifurcation of society, eventually resulting in an unskilled lower class and a skilled upper class.

In the second part, The Current Global Environment Is Unique, I established the circumstances that characterized the global economic environment at the time. Briefly,

• The economy was global in nature, so that actions undertaken by market participants (businesses, policymakers, etc.) would have global impacts on other participants.

• I wrote that the global economy was in recession, however I was wrong on that point. The global economy had largely recovered from the Great Recession, but the economic environment was still relatively fragile.

• Society was in a long-term state of transition as it adapted to ongoing introductions of new technologies.

In the third part, The Nature of Jobs Has Changed over Time with the Introduction of New Technologies, I compared jobs counts by detailed occupation categories in 1940 to those in 2013, and I established that over the 73-year period, new technologies had:

• Eliminated many jobs by substituting technology for human labor,

• Created many new jobs by creating demand for new types of human labor, and

• Changed the nature of many jobs by complementing human labor.

In the fourth part, What Automation Can and Cannot Do, I established that computers could perform any cognitive or manual tasks that could be clearly defined and specified by a set of rules. On the other hand, computers could not perform tasks for which data inputs and/or outputs were not well-defined or could not be fully pre-specified. Such tasks involved

• Solving unstructured problems,

• Working with new Information, or

• Performing non-routine manual tasks.

Alternatively described, computers did not possess the ability to

• Be creative or form new ideas,

• Recognize large-frame patterns,

• Engage in or with complex communications, or

• Perform tasks requiring sensorimotor skills.

The fifth part of the analysis, What Will the Future Hold for Jobs?, was my general discussion and conclusions.

One of the more important things I learned during my analysis was that there’s a world of difference between tasks and jobs. Jobs involve people performing many different tasks. And while automation may very well eliminate the need to perform many tasks, that will simply change the nature of the jobs people perform around various automated tasks.

Of course, I would recommend you read my reasoning. However, the conclusion I came to, which I believe holds equally well today, is this:

So, what we have is that the global nature of the world’s economies suggests that competition will continue to increase the pace of new technology development, while prices in existing markets will continue to decrease and whole new markets will continue to appear (this is the premise of Brynjolfsson and McAfee’s The Second Machine Age). The pace of new technology adoption in different markets will certainly be delayed by such factors as consumer resistance or other market frictions (discussed in Part 2). However, technology will eventually prevail. In most cases, experience has shown that new technologies are more likely to complement humans and change the nature of their jobs than it is to replace workers entirely, though in some cases jobs will certainly disappear. This should continue to be true, unless or until new technologies are able to master inherently human tasks or skills. Given the continued discovery of new complexities in the human body – and especially in the brain – artificial intelligence is not expected to be able to master human skills anytime soon.

A combination of improvements in technology and increases in standards of living in developing nations should lead all workers globally, skilled and unskilled alike, to be better off. Mark P. Mills in “In The Future, Will Only Robots Celebrate Labor Day?” sums it all up beautifully:

Since 1890, we have seen profound technology changes; and we have come to take them for granted. It’s easy to forget that 1890 was literally, not figuratively, the horse-and-buggy era.

Such technological progress drove productivity, boosted the economy, and created jobs – and new kinds of jobs. No one in 1914, when half of America lived on farms, imagined the kinds of jobs that would exist by 1964 – for example, those associated with radio and television and electronic media. Similarly, no one in 1964 imagined the companies and jobs that would emerge by 2014 – for example, the $300B software industry employing multitudes. The U.S. Census didn’t even have a job category called “software” in 1964.

Moreover, only a fraction of those employed in software companies are coders; just as only a fraction of those employed in radio and television were and are employed in electrical engineering. In all sectors, whether software, information technology, or shoes, most employees are needed for office functions, management, sales, and support services common in all businesses.

…

In order to ameliorate the inevitable disruption that progress brings, policymakers should get out of the guessing game and let the millions of varied private businesses take primary responsibility for both creating new jobs and re-training employees as needed. The government, for its part, should stick to what is knowable about the drivers of job growth. Job creation happens when economies boom as result of an expanding private sector, and especially with the proliferation of small businesses, start-ups and entrepreneurs. It should go without saying, but it does require saying, that the key thing government can do is to create conditions that accelerate private sector growth and new business formation.