The following contract offered on the prediction market Manifold neatly encapsulates pessimism about how artificial intelligence will affect US workers: “If the U.S. unemployment doesn’t exceed 30% due to AI before 2040, which of these will be major reasons why?” According to the available options, avoiding mass unemployment would depend mostly on the technology falling short or on bottlenecks and political pushback slowing it down. But something is missing from the options in the graphic below:
Actually, two options are missing—ones I’m sure plenty of amateur forecasters would like to be able to choose from, especially those with some knowledge of economic history. After all, we’ve had lots of technological progress over the past 250 years, and yet work hasn’t vanished.
The first missing option involves demand. Technological progress can make workers more productive and allow businesses to lower prices for goods and services. In that case, consumers might buy more of those goods and services, supporting jobs in those industries.
They might also take the savings and spend them on other goods and services, which would support jobs elsewhere.
The second missing option: AI can create new tasks, products, industries, and jobs. This may seem like an obvious impact, but it’s one too often ignored because it’s extremely difficult to predict what all those new things will be exactly. Still, the general possibility should be incorporated into baseline thinking about how AI can affect jobs. A good example is this March 2023 analysis from Goldman Sachs, “The Potentially Large Effects of Artificial Intelligence on Economic Growth”:
We anticipate that many workers that are displaced by AI automation will eventually become reemployed—and therefore boost total output—in new occupations that emerge either directly from AI adoption or in response to the higher level of aggregate and labor demand generated by the productivity boost from non-displaced workers. Both of these channels have plenty of historical precedent. For example, information technology innovations introduced new occupations like webpage designers, software developers, and digital marketing professionals, but also increased aggregate income and indirectly drove demand for service sector workers in industries like health care, education, and food services.
A few other things worth pointing out here: Massive, AI-driven unemployment in coming years isn’t the baseline of most technologists or economists. Despite the supposed increase in AI capabilities, we’re still not seeing much labor-market impact. And, finally, if AI is an important general-purpose technology, there will be labor market disruption. But that can happen within the sort of dynamic, churning labor market that’s familiar to workers. I mean, that’s my forecast.
