Study: AI Could Drive 20-30% U.S. Unemployment Soon

Verizon CEO Dan Schulman made waves in an April 2026 interview with The Wall Street Journal when he predicted that AI and robotics could push U.S. unemployment to 20%-30% within two to five years. Schulman, who took the helm at Verizon in October 2025, has positioned himself as one of the most outspoken corporate leaders on AI's disruptive potential, urging fellow CEOs to be transparent with employees about looming job losses — particularly in manual labor roles threatened by advances in humanoid robotics. His warnings carry added weight given his own actions: just one month into his tenure, he oversaw the layoff of 13,000 Verizon employees while launching a $20 million retraining program to help displaced workers transition to new roles.

The prediction has struck a nerve with the public. A Quinnipiac University survey found that 7 out of 10 Americans are already concerned that AI will lead to job losses, and research from the Federal Reserve Bank of St. Louis has identified a correlation between AI prevalence and rising unemployment in highly exposed occupations between 2022 and 2025. A Boston Consulting Group report estimates that 50%-55% of all U.S. jobs will be impacted by AI, with up to 15% potentially eliminated entirely.

However, many economists and researchers argue Schulman's figures are far more dramatic than the evidence supports. Goldman Sachs Research projects a more modest 0.6 percentage point increase in unemployment over a 10-year transition period, while a Yale Budget Lab study found no clear evidence that AI exposure is currently driving changes in overall employment levels. The debate underscores a fundamental tension: whether AI will be a net job destroyer on an unprecedented scale, or whether — as with past technological revolutions — it will ultimately create as many opportunities as it displaces.