A new analysis from the Brookings Institution challenges a widely held belief in the AI age: that worker retraining can meaningfully counteract the large-scale job displacement caused by artificial intelligence.

The piece, “AI, labor displacement, and the limits of worker retraining”, argues that while reskilling efforts are well-intentioned, they fall far short of what’s needed to address the scale and speed of disruption now underway.

The analysis, written Google DeepMind AI Policy Researcher Julian Jacobs, draws on decades of labor market data and policy outcomes. Muro points out that despite ongoing political rhetoric promoting retraining as a panacea for technological disruption, historical evidence shows these programs have had limited success.

“In truth,” Jacobs writes, “the country has never shown much ability to deliver effective worker retraining in the wake of large-scale economic shocks.”

The report underscores that AI is no longer confined to automating repetitive or low-wage labor. It is now displacing mid-skill and even high-skill white-collar roles. That expansion of risk makes traditional retraining efforts—often underfunded, poorly targeted, and slow-moving—insufficient for the urgency of the moment.

Rather than relying on reskilling as a cure-all, Jacobs advocates for a broader policy response. This includes slowing the pace of displacement through regulation, boosting labor mobility, investing in job creation, and strengthening social safety nets.

“It’s time for a more honest reckoning,” he writes, urging policymakers to stop overselling retraining and start developing structural solutions to an AI-transformed economy.