Robot Overlords and Red Herrings

A few days ago, a single blog post tanked the stock market. I had to check it out.

It’s about how AI will crash the economy in a few years. It’s catnip for those of us who like big-picture writing: It’s ambitious, wide-ranging and provocative, and it’s written with a bit of attitude. If you haven’t seen it yet, it’s worth setting aside 20 minutes or so to get through it.

The core of the argument is that agentic AI will replace human labor in most white-collar industries and will do so with dizzying speed. The consequent abrupt and massive job displacements will lead to crashes in property values and local tax bases, with devastating impacts on communities and much of the public sector. Mass unemployment will occur simultaneously with massive upward concentration of wealth (and therefore power) to those who own the technology, so the government won’t be in a position to help matters much; it will have been sufficiently captured by those with means that everyone else will be reduced to spectators. Loss of buying power among millions of people will lead to cascading business failures as most people are forced to cut spending and those who haven’t lost their jobs yet will see what’s coming and hoard what they can, while they can.

Anyone familiar with Marx and Keynes will recognize much of the argument. Capital-intensive production relies on mass consumption to buy what it produces; displacing the workers (in this case, in favor of bots) will collapse aggregate demand. Marx called that a contradiction, though his solution left much to be desired. Keynes offered a solution to slack demand—deficit spending to “prime the pump”—premised on the idea that recessions are brief and cyclical. But the changes the bots will force—massive disintermediation, basically—will be structural and permanent. Short-term stimulus isn’t a long-term solution.

The new element in the argument is the bots themselves. If they can do the things that we used to have to pay smart people to do, what’s the economic value of smart people? Yes, you need a few to run the bots and some skilled folks to tend to the physical infrastructure, but that doesn’t come anywhere close to full employment. An economy built on the labor theory of value will struggle when you don’t need labor to create value anymore. What to do with all those people?

Ahem.

Decades of austerity politics have drained our political imaginations to the point that an entire school of thought has been forgotten. Over the first two-thirds of the 20th century, it was a commonplace assumption that the workweek would get significantly shorter over time. Why wouldn’t it? Massive productivity gains offered the prospect of greater choice in how to spend our time. Splitting the gains between increased wages and increased free time only made sense. What we now call, wistfully, “work-life balance” looked for a while like it would take care of itself. (Juliet Schor’s classic The Overworked American is very good on this.)

By the 1960s, the assumption of having solved scarcity drove much social thought. John Kenneth Galbraith’s The Affluent Society—now remembered mostly for coining the term “conventional wisdom”—set the key dilemma for modern market economies as creating enough demand to consume everything that modern technology could produce so easily. (That was the social function of advertising.) The sociologist David Riesman titled his 1960s collection of essays Abundance for What?, confident in the assumption that scarcity was no longer relevant in advanced countries. Now we could turn our attention to what he called the “cultivation of leisure.”

Contrast Abundance for What? with last year’s bestseller Abundance, in which Ezra Klein argued that progressives had to convince themselves that there was even an alternative to austerity. The Riesman of the 1960s would have found the premise preposterous. Productivity is so much higher now than it was even in the ’60s—why aren’t we awash in wealth? Why are our public institutions gasping for air when the economy is more productive than it has ever been? Why is the workweek getting longer? Why is daily life getting harder?

It’s not about assembly lines, mass media or bots. It’s about a more basic political choice: Should the fruits of technological advance be hoarded exclusively at the top, or should they be shared broadly through the polity? Should the people serve the economy, or should the economy serve the people?

Robot overlords are red herrings. The key issue is power. Get that right and the bots may serve a useful purpose. Get that wrong and the viral memo may prove prophetic. We should choose while we still can. That’s the current value of smart people.