AI is currently eliminating office jobs and reducing salaries

AI's Economic Impact: Rethinking the White-Collar Future

In February, a leading financial newsletter on Substack sparked intense debate with a provocative scenario: What if the surge in artificial intelligence, which has already generated immense wealth and driven unprecedented corporate investment, signals not a bullish bubble but a looming downturn? Could the very technology boosting office productivity ultimately undermine the broader white-collar workforce?

A imagined a future memo dated June 30, 2028: “The unemployment rate hit 10.2% this morning, a 0.3% unexpected increase. Markets dropped 2% on the news, pushing the S&P 38% below its October 2026 peak.”

This thought experiment had immediate real-world consequences. The Dow Jones fell 1.7% the following Monday. Companies mentioned in the post, such as Monday.com and DoorDash, each lost about 7% of their value, while IBM plunged nearly 13%.

In essence, a speculative Substack article triggered a multi-billion-dollar market loss. This reaction raises a key question: Did the post tap into deep-seated anxieties about AI, or was it simply an overreaction to an unlikely scenario?

The Ongoing Decline in White-Collar Employment

The question posed by Citrini—what happens when an economy built on valuing human intelligence faces a world where that value diminishes—is increasingly on the minds of economists, labor experts, and workers. While comprehensive data is still emerging, early indicators are concerning.

White-collar employment has now shrunk for 29 straight months. Aaron Terrazas, a former chief economist at Glassdoor, calls this unprecedented. “White-collar hiring has slowed dramatically, and payrolls have contracted in a way we haven’t seen outside of recessions in the past 70 or 80 years,” he explained. “This should be raising serious concerns.”

However, the overall unemployment rate—still near 4.3%—masks these white-collar challenges. Terrazas notes that this figure has become less reliable, as underemployment and workforce exits become more common than outright unemployment. He points to declining job postings and hiring rates as more telling signs. “We’re seeing warning signs across various sectors,” he said.

AI's Disruption: More Than Just Job Losses

Daniel Keum, a Columbia Business School professor specializing in AI’s workplace effects, is direct: AI is reducing demand for white-collar employees. He describes the current period as a “technological shock” with two phases.

• The first phase is already underway: AI is replacing office workers in the U.S., not just enhancing their productivity.
• “Labor costs in the U.S. are high,” Keum explained. “AI is being used to cut staff and reduce expenses, which yields significant savings.”
• The second phase—AI-driven growth through new products, services, and jobs—may still be years away. For now, companies are experiencing cost reductions without the offsetting benefits of new revenue streams.

Not all job losses are direct replacements by AI. Some employees are let go because companies are shifting resources toward AI initiatives and away from other areas. Meanwhile, the massive investments by tech giants like Amazon, Microsoft, Google, and Meta are funding data centers, not hiring more people.

Keum argues that soaring business investment doesn’t necessarily mean more jobs for degree-holders—in fact, it may signal the opposite.

He suggests tracking the job market for new MBA graduates as a leading indicator. If even the most sought-after candidates are struggling, it signals a shift in elite labor demand.

The numbers are troubling. In January, The Wall Street Journal reported that 21% of job-seeking graduates from Duke’s Fuqua School of Business were still unemployed three months after graduation in 2023, up from 5% in 2019. Georgetown’s McDonough School saw 25% still searching, compared to 8% in 2019. At Michigan’s Ross School, the figure rose to 15% from 4%. Even Harvard Business School had 16% of graduates unemployed after three months, higher than pre-pandemic levels.

Other factors are also at play, including changes in immigration policy, post-pandemic hiring adjustments, high interest rates, and unpredictable trade policies. Still, the weakening demand for highly credentialed workers is striking. If top business schools are seeing more graduates struggle to find work, something fundamental has shifted.

White-Collar Wage Pressures: Subtle but Significant

The Citrini post also explored the risk of declining white-collar wages—a trend Keum is monitoring closely. Historically, pay has tracked closely with productivity, but as AI boosts output, workers are finding it harder to capture the extra value they generate due to reduced demand for their labor.

When automation can replace your work, your bargaining power drops. “A junior law associate might have previously earned 20% of billable hours,” Keum noted. “Now, they might bill more hours but only receive 10%—because AI is always an alternative.”

If AI continues to erode workers’ ability to claim a larger share of value, it could accelerate a long-term trend. In the U.S., labor’s share of GDP—a measure of how much value goes to workers versus capital—has been declining for decades, dropping nearly 10 percentage points since its peak in the late 1960s and early 1970s, reaching 56% in 2024.

It’s not always easy to see if white-collar pay is falling, partly because detailed data is scarce and salaries are often “sticky.” Companies rarely cut base pay directly, but they may reduce compensation in less obvious ways.

• Benefits packages may quietly shrink, such as employers covering less of health insurance premiums.
• Non-salary compensation, like stock grants or bonuses, may be reduced.
• Job responsibilities may increase without a corresponding pay raise—a phenomenon Terrazas likens to “shrinkflation.”

These changes can erode overall compensation, even if salaries remain unchanged. Recent data from Sequoia shows that the percentage of companies fully covering employee-only health premiums has declined for three consecutive years.

Even if these adjustments don’t show up in headline wage statistics, they still reduce take-home pay.

Potential Ripple Effects Across the Economy

Citrini’s scenario painted a bleak picture of widespread white-collar layoffs and diminished earning power rippling through the broader economy—turning prime mortgages into risky loans, shrinking demand for everything from cars to vacations and private schools. The resulting “consumption hit” could be far greater than the number of jobs lost alone would suggest.

Yet, Terrazas cautions against panic: “So far, the evidence points to moderate, not seismic, changes, and there’s no definitive proof that AI is the main cause—just a lot of warning signs. Since data always looks backward, it may just be a matter of time. The scenario described would be unprecedented, but sometimes history doesn’t repeat itself.”

He adds, “Most people agree that workers will adapt—and are already adapting—to these labor market changes. The real question is whether they’ll end up better or worse off. Citrini’s authors assume adaptation means taking lower-paying or less prestigious jobs, but that’s not always the case.”

Not everyone believes alarm is justified. Some top economic officials have pushed back, with Federal Reserve Governor Christopher Waller stating, “AI is a tool. It’s not going to replace us as human beings. This is somewhat exaggerated.” Historically, every wave of automation has eventually created more jobs than it eliminated.

However, that optimism rests on the assumption that new roles will always require human workers—an assumption now in doubt. Previous technologies replaced tasks, but not the need for human creativity and judgment. The future may not mirror the past.

The Citrini post resonated because it suggested that this time might truly be different. While some believe white-collar workers will ultimately adapt and thrive, current trends suggest their influence and bargaining power may be diminishing, not growing.