Great Wisdom On AI And The Future Of Work: We’ll Figure It Out

Jobs and artificial intelligence continue to be a hot topic two years after I first wrote about job losses and the net gains from AI. Dario Amodei (a co-founder of Anthropic) recently made headlines on the topic, then the Society for Human Resource Management (SHRM) issued a report, which led to insightful comments from Christopher Lind and a short but very wise observation by Katherine Mangu-Ward.

Amodei sounded a loud alarm in an interview with Axios. “AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10-20% in the next one to five years, Amodei told us….” The news article continues, “Amodei said AI companies and government need to stop “sugar-coating” what’s coming: the possible mass elimination of jobs across technology, finance, law, consulting and other white-collar professions, especially entry-level gigs.” (Disclosure: I have a close relative who works at Anthropic.)

The SHRM report might calm the worst fears, finding “12.6% of current U.S. employment (19.2 million jobs) faces high or very high automation displacement risk.” That’s probably only calming after reading Amodei’s comments; it’s still a big harm. Less calming is the authors’ explanation of how they arrived at that figure: “… the methodology we adopt makes the central assumption that an occupation’s current automation level is indicative of future automation displacement risk …” That may not be the worst possible assumption, but could an occupation be clobbered by technology not currently in use? Or could the greatest changes be in those occupations that currently are least automated? Thinking through that question leaves me less certain than before I read their report.

Christopher Lind used the SHRM report first to tell us to slow down. In a video he said that there is a difference between the risk of change and actual implementation. Although many talk about AI, actual usage in everyday work is less common. But he does believe that change is coming, if not tomorrow. In his Substack he wrote, “… most people won’t wake up and find their job is gone, despite what headlines indicate. Most will slowly discover their job is no longer what it used to be.”

Let’s roll together what we know. Productivity has increased in some occupations, enabling more output for a given number of hours of work. So far the largest implementations have been for computer programmers and customer service representatives. Programmers finish projects much faster by having AI models write code at the programmer’s direction. Customer service representatives are improving their answers, leading to faster and more accurate responses. Improved productivity means, at first, lower demand for that type of labor. In a supply-and-demand exercise, the demand for labor shifts to a lower quantity demanded at any wage. Positions that are open due to normal employee turnover go unfilled, and then layoffs occur.

The people who remain have been earning wages that are now above market, so they go without raises. When new hires are needed, they are offered lower wages because so many people in the field are unemployed. It’s easy for a company to hire someone, and low-ball job offers are accepted.

Now that a given task is less expensive, the company chooses to have more tasks done. That will entail fewer employees than before the productivity improvement occurred, but more than if wage rates had not fallen. In practical terms, we’ll eventually see web sites that are more user-friendly, and customer service calls handled more promptly.

But the economy will be better than it looks at first blush. Greater productivity leads to lower cost of production. Competition among producers leads to lower prices to consumers. (This gets wrapped up in overall inflation policy, so it’s better to think of consumer prices in terms of hours of labor. A car might go from 1500 hours of average work to 1400 hours.) In the aggregate, the benefits of improved productivity are passed on to consumers.

Or will the developers of AI reap all of the rewards? Some people, including Amodei, have voiced concern about AI leading to more for the “haves” and less for the “have nots.” But innovators typically reap just two percent of the value they produce, according to research by Nobel laureate economist William Nordhaus. The rest goes to consumers and workers. One avenue through which consumers will gain will be new companies developing AI models. Although the major AI companies today have some proprietary techniques, the bulk of their methods are widely known. With enough resources and some time, a team of bright people could develop another pretty good AI model. And if AI companies are reaping large rewards from their service, then start-ups will surge to take their share—with lower prices.

We still have the question of how the economy will adjust. Just where will the new jobs emerge, how will people train for the new jobs, or train for the old jobs that have been changed by AI? The greatest wisdom on this question comes from Katherine Mangu Ward on The Reason Roundtable (at about 12:28 minutes in): “We’ll figure it out.”

That may not satisfy the worriers among us, but it is precisely how humans have dealt with change in the past. Entrepreneurs see opportunities when things change. Some of the people are experienced entrepreneurs; other become entrepreneurs when they see the opportunity. With AI will come different costs for different products and activities. There will be increased demand for some things, with certain resources cheaper. Entrepreneurs will step in, form new companies or transform old ones, and create work opportunities.

Talent will not be wasted, for we are far, far away from everyone having everything that they could possibly want. Some people will work minimal hours for a simple lifestyle—simple by modern standards, luxurious by historical standards. Others will find 40 hours of work every week for a pretty nice standard of living. Yet others will work most of their waking hours, motivated by both enjoyment of work and the luxuries that work will buy. In short, we’ll figure it out.

It is rare for anyone to predict the exact way that people adjust to new technology. But figure it out is what we do.