Global anxiety over Artificial Intelligence (AI) replacing human workers deepened this week after Amazon announced a major round of corporate layoffs, citing the need to reorganize operations around AI technologies. Economists, however, remain divided on whether automation is truly to blame for the wave of job cuts sweeping through the US tech industry.
Amazon’s decision to trim about 14,000 corporate positions follows similar announcements from Chegg, Salesforce, and UPS — all of which partly attributed workforce reductions to the use of AI tools and machine learning systems. Yet researchers argue that linking every layoff to automation oversimplifies what may be a normal economic correction across industries.
Analysts caution against AI panic
“There is a real tendency to overreact to individual company announcements,” said Martha Gimbel, Executive Director at Yale University’s Budget Lab. “Many of these firms expanded too rapidly during the pandemic and are now scaling back in line with the broader economic cycle.”
Chegg recently cut almost half of its staff, citing “new realities” created by AI chatbots disrupting online education. Salesforce eliminated 4,000 customer service roles, saying AI assistants now handle much of the administrative workload. UPS has shed 48,000 jobs since last year, with management linking the changes partly to automation in logistics and route planning.
Experts note that many of these announcements coincided with the Federal Reserve’s interest rate hikes, which began soon after the launch of ChatGPT in late 2022. The timing, they say, suggests that macroeconomic forces may be as influential as technological change.
Cyclical correction or automation shift?
Data from the Federal Reserve Bank of St. Louis shows that unemployment has risen more sharply in jobs more exposed to AI. However, a study from the University of Pittsburgh found that the only occupations immediately affected by ChatGPT’s launch were administrative and clerical roles, not those in computing or mathematics.
“Both tech workers and admin staff are in a tougher job market than two years ago,” said Professor Morgan Frank, who led the study. “But I’d be skeptical that AI is the sole reason for all of it.”
Enrico Moretti, an economist at the University of California, Berkeley, said large firms like Amazon are “at the forefront of both producing and consuming AI.” While automation may explain part of the company’s job cuts, he noted that many layoffs reflect a correction after years of overhiring during the pandemic.
Hiring, firing, and the business cycle
Amazon’s quarterly results in July beat Wall Street expectations, with sales rising 13 percent year-over-year to $167.7 billion. Analysts say the company’s strong performance indicates reorganization rather than distress.
Lawrence Schmidt, a finance professor at MIT Sloan School of Management, said automation increasingly shapes hiring decisions. “It’s not surprising if Amazon chooses not to hire in areas that can be automated quickly,” he said. “Even if total employment remains steady, we’ll likely see reallocation rather than elimination.”
Gimbel agreed that most workforce changes still align with traditional business patterns. “So far, nothing suggests a new trend outside normal hiring and firing,” she said. “The real question is whether AI will permanently alter these patterns once economic growth picks up again.”
Long-term implications
Economists largely agree that distinguishing between cyclical and AI-driven job losses will be key to understanding future trends. If a slowdown hits, human resources and marketing professionals — both heavily exposed to AI — would likely face early job cuts. Yet such trends, they note, resemble historic recessions more than a technological revolution.
While AI may eventually reshape the world of work, analysts believe the current wave of layoffs represents a broader corporate recalibration after years of aggressive expansion. “AI is part of the story,” said Gimbel, “but it’s not the whole story.”
The report was originally published by the BBC.
