White collar workers displaced by AI could spark a revolution

Earlier this month, Microsoft announced plans to lay off 9,000 employees, joining the ranks of tech companies cutting headcounts while touting the productivity gains enabled by artificial intelligence.

According to CEO Satya Nadella, up to 30% of the company’s code is now being written by AI. Salesforce CEO Marc Benioff one-upped him after his own firm’s layoffs, claiming that up to 50% of their work is now being shouldered by AI.

These are not simply corporate earnings call updates. They are previewing the future of work in America.

Claims from leaders at OpenAI and Anthropic that their products will change the world deserve some skepticism. After all, snake oil salesmen project confidence, too.

But when layoffs start happening and the heads of publicly traded companies tell you why, it’s worth listening.

Displacement’s here

The question isn’t whether AI will displace workers, but which workers — and what happens when America’s comfortable laptop class finds itself on the wrong side of technological disruption for the first time.

Over the past 50 years, Americans were sold on the promise of globalization and the idea that we’d lead the world in knowledge work and no longer have to make things ourselves. We’d “learn to code” and someone halfway around the world would assemble our iPhones.

The most searched topics on Anthropic’s AI platform Claude. Merrill Sherman / NY Post Design

College-for-all became the guiding principle of education policy, and the dollars followed: Since 1980, public spending on higher ed has more than doubled.

Meanwhile, career and technical education — the kinds of programs that lead to good blue-collar jobs after high school — has been ignored, with its share of federal K–12 spending dropping to low single digits.

In other words, America has systematically overinvested in precisely the kinds of jobs AI seems primed to disrupt.

Right now, these disruptions seem largely focused in the tech sector. But that’s likely a harbinger of things to come.

The history of innovation shows repeatedly that when general-purpose technologies are introduced, they take time to diffuse throughout the economy — and the first place we see disruption is in their own industry.

Electricity is a perfect example. As the economic historian Paul David shows, the first productivity bump from electrification was confined to the power plants themselves — they learned to squeeze more kilowatt-hours from each ton of coal and worker.

Only gradually, over the following decades, did factories reorganize around electric motors, and only then were these gains mirrored across the economy.

In computing, too, David shows how the initial gains were seen in the computer manufacturing sector, followed by sectors that were already using early computers the most. Eventually, every company from Walmart to Bank of America had its own IT department.

Elite ‘overproduction’

AI seems to be in the early stages of following this precise pattern, starting on its home turf of software. It’s one thing for software engineers to be concerned for their futures, quite another for mass layoffs to hit economy-wide.

Just as electrification ultimately came for agriculture and manufacturing, AI will come for law, and banking, and marketing.

When the last big economic disruption of globalization struck blue-collar workers, sending manufacturing jobs overseas, the result was hollowed-out communities, an opioid crisis and an unfulfilled promise of “retraining” for “the jobs of the future.”

Now, the AI wave is primed to hit those exact jobs, and the people who built their own prosperity on the wave of globalization that left so many others behind.

But this isn’t just a mirror image of the globalization crisis.

On the one hand, high-income, highly educated elites have much greater resources at their disposal to weather an economic storm.

On the other, the power elites hold may pose a far greater risk to political and social stability when they find themselves under threat.

The complexity scientist Peter Turchin has shown how, throughout history, an excess of elites with poor economic prospects is the path to political instability. Going back as far as the late Roman Republic, elite overproduction is a key contributor to chaos (alongside stagnating or declining general living standards and increased public debt — which may sound familiar).

Disruption factor

The American laptop class is a political powder keg. As the comfortable jobs they were promised become harder to land, and the self-validating stories they told themselves of their own value come unwound, competition will become fierce.

As more and more of them end up on the losing side, Occupy Wall Street may look like a mild preview of what’s to come.

How long will this all take? If we have decades to adapt, older workers will have time to learn new skills while younger workers can adjust their own career paths. New industries can spring forward to use the labor no longer dedicated to litigating in court, speculating in markets and posting on social media.

But if the disruption happens more quickly, we risk creating millions of displaced knowledge workers who will have not only the skills and networks, but also the time and motivation to organize a radical political movement.

Predictions about what’s to come are mostly educated guesses, drawing on history and an expectation that the lines on the graph will continue to go up.

This is not enough.

Policymakers cannot neither rest comfortably in Sam Altman’s promises of utopia, nor overreact to overwrought warnings with regulation that short-circuits valuable innovation that may take decades to deploy, nor simply wait and see what happens.

It’s time to prepare

The Bureau of Labor Statistics should begin to track AI displacement in real-time, sector by sector, occupation by occupation.

Instead of waiting for quarterly employment reports that tell us what happened months ago, live data should highlight which jobs are disappearing and where new ones are emerging.

Anthropic has already begun leading the way on this type of research, reporting how its tools are being used within each of the BLS’s job categories.

AI has the potential to remake the American economy — it may already have started.

The last century was defined by the remarkable gains of technological progress as electrification and then the digital revolution washed across society, and the shocking costs of foolish enthusiasm for the rapid disruptions of globalization and deindustrialization.

The next will be defined in part by which of these transitions AI resembles.

The choice isn’t whether AI will displace knowledge workers — it’s whether we’ll see the disruption coming and prepare accordingly, or wait for the laptop class to take to the streets.

Abigail Ball is the executive director of American Compass.