AI Job Fears Echo 1931: What History Says About Change

In December 1931, economists gathered in Washington as the Great Depression ground on, asking a question that sounds strikingly current: Is technology destroying American jobs?

The examples were concrete. Dial telephones cut work for switchboard operators. “Talkies” left movie-theater musicians with fewer jobs. Gasoline-powered trucks displaced teamsters who drove horse-drawn teams. Those were real livelihoods, and families felt the loss.

Ninety-five years have passed since that meeting. Now a new generation of economists is asking the same questions about artificial intelligence.

The 1931 Debate

The economists at the gathering didn’t pretend technological disruption was new. They said it wasn’t. What worried them was the pace. Sumner H. Slichter, a Harvard economist, told colleagues at the annual meeting of the American Economic Association that change was happening “more rapidly than in any previous age.”

He went further. “And yet we have failed, for the most part, to recognize that change presents a major social problem,” Slichter said.

The economists warned that the speed of change, and its effects on the economy and social stability, posed particular risks. That was their argument in 1931, as the country reeled from economic collapse.

The Same Questions, New Jobs

The occupations under scrutiny have changed. Today’s economists are examining computer programmers instead of telephone operators, scriptwriters instead of movie musicians, and drivers of diesel-powered trucks instead of animals.

The questions, though, have barely moved. Will technology’s benefits reach broadly, or will they mainly accrue to the wealthy? Will new jobs emerge in time to replace those lost? Can government data measure an economy changing this quickly? And what, if anything, can the government do to slow disruption or help affected workers?

That last question deserves scrutiny. It carries an assumption: Washington has the tools to slow technological change or cushion its effects without making matters worse. In 1931, economists met during a depression, when the government had limited capacity to read the economy. Today’s economists are asking whether government data can keep pace with an economy moving faster still.

What Measurement Can and Can’t Do

The question about data matters more than it might seem. Any policy response depends on knowing what’s actually happening. When official statistics lag behind a fast-changing economy, policymakers risk acting on yesterday’s picture of today’s problem.

The broader question—whether new jobs will emerge in time—is just as important. The article’s account of the 1931 debate doesn’t answer it either way. Instead, it shows that economists have wrestled with the same worry for nearly a century, across telephones, talking pictures and trucks, while each generation faced uncertainty about how quickly workers and employers would adjust.

Concerns about who benefits are legitimate and persistent. So is the case for humility about what government can accomplish when it tries to slow innovation already underway.

Why This Matters:

The 1931 meeting reminds us that technological disruption has long prompted calls for government action. The record in the source shows economists then and now asking whether Washington can slow change or help displaced workers, and whether its own data can track the economy in real time. For taxpayers and employers, those aren’t academic points. Policies built on incomplete measurement can carry real fiscal costs and unintended consequences. Markets and workers’ own adaptability have resolved, in the past, the core uncertainty over whether new jobs will arrive in time to replace lost ones. Any government response to AI will have to justify itself against that history and the limits of what officials can know.