Lawmakers want better data on AI’s workforce impacts

Summary

A bipartisan pair of senators is pushing for more and better data to understand the impact of artificial intelligence on the workforce. That data could help back up what experts told a Senate subcommittee on Wednesday — that the biggest AI-fueled change may be different tasks at work rather than mass unemployment. Banks introduced the bill, known as the AI Workforce Projections, Research and Evaluations to Promote AI Readiness and Employment, or AI Workforce PREPARE Act, in December. At a hearing of the Senate Health, Education, Labor and Pensions Committee’s panel on Employment and Workforce Safety, which Banks chairs, he said that better data is a first step toward helping workers master skills needed “in the AI economy.” “Existing labor statistics too often fail to tell us how occupations’ tasks are changing, what skills are needed, which jobs will grow and shrink and how workers move through the labor force because of AI,” Banks said. The bill would allow the Labor Department to hire experts in AI and require the department to set up an AI Workforce Research Hub, monitor how employees move between jobs impacted by AI and develop benchmarks to measure which tasks may be automated, requiring employees to be retrained. It would require employers to disclose if AI was a factor in a mass layoff. Maggie Hassan, D-N.H., Jon Husted, R-Ohio, and Roger Marshall, R-Kan., are also co-sponsors of the legislation. Hickenlooper said the bill would increase communication about forecasts for workforce needs in the age of AI, as well as strategies to respond. Liya Palagashvili, director of the Labor Policy Project at the Mercatus think tank at George Mason University, said that measurement systems should be “resilient across several possible futures.” She said that data on AI in the labor market could be improved by “linking business surveys’ answers to actual hiring and wage outcomes” to allow comparisons based on whether a firm uses AI. Palagashvili said that “early evidence does not yet show broad AI-driven employment loss.” She said that current data on jobs for young workers in fields most exposed to AI is “more consistent with slower hiring than with layoffs” and that some of the slowdown may have begun before the release of major generative AI models, indicating that the hiring changes could be caused by other economic factors.