Azernews.Az

Akbar Novruz

Read more

Economists who once dismissed concerns about artificial intelligence disrupting the labor market are beginning to change their tone, warning that while large-scale shocks have not yet materialized, the risk is growing, and policymakers may not be prepared, AzerNEWS reports.

According to the New York Times' article, for years, skepticism dominated economic discussions around AI. Rising unemployment among university graduates was largely attributed to high interest rates and macroeconomic uncertainty, while warnings of mass layoffs were often dismissed as misunderstandings of past technological transitions. Even job cuts linked to AI adoption were frequently labeled as “AI-washing,” with companies accused of masking broader business failures.

But a subtle yet significant shift is now underway.

Many economists still see limited evidence that AI has already destabilized labor markets or fundamentally transformed productivity. However, an increasing number believe such disruption could soon emerge.

“I don’t think AI has yet delivered a major blow to the labor market or dramatically changed corporate productivity,” said Daniel Rock of the University of Pennsylvania. “But I think it’s only a matter of time.”

A recent survey asking economists to project the next five and 25 years found that most expect modest acceleration in economic growth as AI advances, broadly in line with historical trends. However, a faster-than-expected technological leap—while less likely—could lead to more dramatic outcomes, including rapid growth, rising inequality, and widespread job displacement.

“Economists are taking AI seriously now,” said Ezra Karger of the Federal Reserve Bank of Chicago.

These views increasingly align with those in the AI industry, where uncertainty dominates. Experts agree that AI could either eliminate entire professions or have more limited effects. Its impact may first hit entry-level office jobs or extend to experienced professionals and even manual labor sectors. The timeline remains unclear, ranging from a few years to several decades.

Given the scale of potential disruption, economists say it is time to begin designing support systems for workers displaced or affected by economic transformation—an area where past technological shifts have often exposed policy shortcomings.

“This issue is now discussed enough that we should begin thinking, as a country, about what policies make sense in a world where labor and career models may change within the next two to five years,” said Robert Seamans of New York University.

When OpenAI launched ChatGPT in November 2022, economist Alex Imas of the University of Chicago was unconvinced it would reshape the economy.

“I understood it was important, but I was initially skeptical,” he recalled.

His view shifted in late 2024, when OpenAI introduced more advanced reasoning models capable of analyzing problems step by step and delivering more reliable answers—significantly expanding potential use cases.

“For me, that was a paradigm shift,” he said. “I thought: this could be something on the scale of the industrial revolution—or even bigger.”

Others point to more recent developments, including tools like Claude developed by Anthropic, and the rise of autonomous AI “agents” capable of completing tasks independently.

Working with such tools, Molly Kinder of the Brookings Institution said she realized that basic research tasks once assigned to students may no longer be necessary.

“Honestly, I’m not sure what a college student could add to my team that Claude couldn’t do,” she noted, adding that higher-level strategic roles remain relatively safe—for now.

“But if your job is entirely computer-based, you may face challenges in the long term,” she warned.

Despite rapid technological progress, economists caution that real economic transformation depends on how quickly companies adopt and integrate AI—historically a slow process.

Legal and regulatory barriers, workforce retraining, and organizational resistance can all delay implementation. As Marta Gimbel of Yale University pointed out, many promising technologies have failed to deliver on their transformative potential.

However, early signs suggest AI adoption may be accelerating. Data from the United States Census Bureau indicates that one in five companies has used AI in the past two weeks.

Some economists believe the impact is already visible. Research from Stanford University shows declining employment in entry-level roles most exposed to AI.

“Usually, productivity gains from technology take decades to appear,” said Erik Brynjolfsson. “This time may be different.”

How distressing will it be?

While some economists foresee significant disruption, their projections remain more cautious than those in Silicon Valley.

Dario Amodei has warned that AI could eliminate up to half of entry-level jobs in the coming years. Investor Vinod Khosla predicts that a vast majority of professions could be automated by 2030, while Elon Musk has suggested work itself may no longer be necessary.

Most economists consider such forecasts exaggerated, focusing instead on the transition period and its challenges.

“The key question is how painful the technological shock will be,” Gimbel said.

Estimates suggest that a large share of jobs could be affected by AI, though not all will disappear. Many roles are expected to evolve rather than be replaced entirely.

“Full replacement is slow because implementation is hard,” said Greg Emerson. “But job transformation happens much faster.”

Still, layoffs in certain sectors appear inevitable, with the severity depending on the speed and scale of change.

Economists increasingly agree that policymakers must act now.

Existing social safety nets, such as unemployment benefits, often fail to cover recent graduates—among the groups most at risk. Retraining programs remain underfunded and frequently ineffective.

Anton Korinek of the University of Virginia warned that this wave of disruption could be more prolonged than previous technological shifts.

He also urged policymakers to consider more radical scenarios, including the possibility that AI could outperform humans across most tasks.

“Our job is to think about the most serious risks—what could happen and how we prepare for it,” he said.

Korinek plans to continue that work not only in academia, but also within the industry, having moved to Anthropic—a sign that the debate over AI’s economic impact is no longer theoretical, but increasingly urgent.