Artificial intelligence is advancing rapidly across industries, and according to JPMorgan economist Murat Tasci, the next wave of its integration could hit white-collar workers the hardest. In a recent note to clients, Tasci warned that AI’s impact might lead to a “jobless recovery” a period when economic growth returns after a downturn, but employment, particularly among office-based knowledge workers, lags far behind.
Tasci’s analysis focuses on “non-routine cognitive occupations” essentially white-collar roles that require problem-solving, analysis, or other specialized mental tasks that historically resisted automation. These jobs include a wide range of office positions in finance, law, marketing, design, and technology. Together, they account for roughly 45% of all household employment in the U.S. That means any prolonged employment slump in this category could have wide-reaching consequences for the broader economy.
Structural Unemployment Risks for Office Workers
The concern isn’t just about job displacement it’s about a structural shift that leaves many workers sidelined for an extended period. Tasci noted that white-collar knowledge workers could face a higher, more persistent unemployment rate if demand for their skills weakens and new positions don’t emerge quickly enough to absorb those displaced by AI tools. “A much larger unemployment risk and anemic recovery prospects for these workers might cause the next labor market downturn to look pretty dismal,” he wrote.
In such a scenario, policymakers might be forced to step in more aggressively, either by easing monetary policy or injecting fiscal stimulus to support workers and speed up job creation. This wouldn’t be unprecedented sectors such as manufacturing and clerical work have already been through similar cycles as automation and globalization reshaped the labor market.
Signs the Shift May Have Already Begun
While the U.S. unemployment rate remains historically low at about 4.2% as of July, JPMorgan sees early signs that AI’s influence is starting to show, particularly among entry-level white-collar roles. Tasci pointed out that the unemployment rate for recent college graduates has risen, which the bank links in part to employers adopting AI tools that reduce the need for junior-level staff.
The share of unemployed workers from “non-routine cognitive” jobs has already surpassed that of “routine” workers those in repetitive sales, clerical, or manual roles in recent years. This marks a notable shift, given that for decades, it was routine work that was most at risk of being automated away.
JPMorgan’s data also shows that routine jobs have steadily declined as a share of total U.S. employment, falling from around 55% in the early 1980s to about 40% today. Workers in both routine cognitive and routine manual roles have already experienced jobless recoveries in past downturns, often taking longer to reenter the workforce as industries adopted labor-saving technologies.
A Cautionary Outlook But Not an Immediate Crisis
Tasci stressed that AI has not yet caused a major shock to employment figures, but the pressures could build over time. Historically, disappearing routine jobs have shown a clear pattern during recessions: each cycle has taken longer to recover from tech-driven job losses. If AI accelerates the same trend in white-collar sectors, the next recession could be different in both scale and shape.
“We are not seeing an imminent downturn in the labor market, though the risks are higher relative to a month ago,” Tasci said. Still, the broader job market has shown signs of cooling. In July, the U.S. added far fewer jobs than expected, and combined downward revisions for May and June erased 258,000 previously reported job gains.
For now, the warning is more about preparation than panic. But as AI capabilities expand into drafting reports, writing code, generating marketing campaigns, and even analyzing legal documents, the competitive landscape for human office workers could change quickly. The coming years may test not only the adaptability of individual workers but also the willingness of policymakers to cushion the transition.
