People have been asking this question for a very long time. The British economist David Ricardo added to the third edition of his classic “Principles of Political Economy”, published in 1821, a chapter entitled “On Machinery”, in which he tried to show how the technologies of Industrial Revolution early could, at least initially, harm workers. Kurt Vonnegut’s 1952 novel “Player Piano” envisioned a near future in America in which theAutomation has eliminated most jobs.
At the level of the economy as a whole, the verdict is clear: so far, machines have not eliminated the need for labor. American workers are nearly five times more productive than they were in the early postwar years, but there has been no long-term upward trend in unemployment.
That being said, technology can eliminate certain types of jobs. In 1948, half a million Americans worked in the coal mines; the vast majority of those jobs were gone by the turn of the 21st century not because we stopped coal mining (the steep decline in coal production, first to natural gas and then to renewables, doesn’t started only about 15 years ago), but because opening up the mining of the pits and removal of the mountain top allowed more and more coal to be mined with far fewer workers.
It is true that the jobs that disappear due to technological progress have generally been replaced by other jobs. But that doesn’t mean the process was painless. Individual workers may not find it easy to change jobs, especially if the new jobs are in different locations. They may find their skills undervalued; in some cases, as in the case of coal, technological change can uproot communities and their way of life.
This type of dislocation has, as I said, been a feature of modern societies for at least two centuries. But something new may be happening.
In the past, jobs replaced by technology typically involved manual labor. Machines have replaced muscles. On the one hand, industrial robots have replaced routine work on assembly lines. On the other hand, there has been a growing demand for knowledge workers, a term coined by management consultant Peter Drucker in 1959 for people dedicated to solving non-repetitive problems. Many people, myself included, have said that we are increasingly becoming a knowledge economy.
But what if machines could support much of what we have historically considered knowledge work?
Last week, research firm OpenAI released, with great excitement in tech circles, a program called ChatGPT, which can carry out what look like natural language conversations. You can ask questions or make requests and get surprisingly clear and even seemingly knowledgeable answers. You can also do some fun stuff: A colleague recently requested and received a secular stagnation analysis in sonnet form, but let’s move on to things that might be financially helpful.
ChatGPT is just the latest example of technology that appears to be capable of performing tasks that not so long ago seemed to require the services of not just human beings, but human beings with an education. substantial formal.
For example, machine translation from one language to another was once a joke; some readers may have heard the apocryphal story of the Russian-English translation program which took “the spirit was willful, but the flesh was weak” and ended with “the vodka was good, but the meat was rotten”. Translation programs these days may not produce great literature, but they are suitable for many purposes. And the same is true in many areas.
You can say that what we often call artificial intelligence is not really intelligence. In fact, it can take a long time before machines can be truly creative or provide deep insight. But then, how creative or deeply insightful is what human beings do? (In fact, how much of what is published in academic journals, a field I know quite well, meets these criteria?)
Thus, a certain number of knowledge jobs can be perfectly replaceable.
What will this mean for the economy?
It is difficult to predict exactly how AI will affect the demand for knowledge workers, as it will likely vary by industry and specific tasks. However, it is possible that, in some cases, AI and automation can perform certain knowledge-based tasks more efficiently than humans, which could reduce the need for some knowledge workers. This could include tasks such as data analysis, research and report writing. However, it should also be noted that AI and automation can also create new job opportunities for knowledge workers, especially in areas related to the development and deployment of AI.
Note that I did not write the paragraph you just read; ChatGPT did, in response to the question “How will AI affect the demand for knowledge workers?” The gift, at least for me, is that I still refuse to use “impact” as a verb. And he didn’t explain explicitly why, in general, we shouldn’t expect an impact on overall employment. But it was probably better than many humans, including some people who imagine themselves intelligent, would have written.
In the long run, productivity gains in knowledge industries, like earlier gains in traditional industries, will enrich society and improve our lives in general (unless Skynet kills us all). But in the long run, we’re all dead, and even before that some of us might find ourselves unemployed or earning much less than we expected, given our expensive education.