Ford School professor and economist Dr. Justin Wolfers recently testified at a hearing for the Michigan House Appropriations Committee. The hearing, titled "Consensus Revenue Estimating Conference," is the first in a series of two hearings determining the Michigan revenue forecasts for the 2025-2026 and 2026-2027 fiscal years.
Much of the public conversation around artificial intelligence (AI) tends to highlight its role as a replacement for human labor, with particular concern for blue-collar and entry-level jobs. However, Wolfers, in his testimony, emphasized a different perspective: he contended that AI will primarily supplement workers in administrative and office-based roles, rather than replace those engaged in physical trades. According to Wolfers, tasks like translation can be completely automated by AI, removing the need for human involvement in these areas. By contrast, plumbers and electricians—whose jobs require manual skills—might use AI to identify issues or streamline their workflow, while the actual hands-on repair work remains in their domain.
"In the past, technology has been a replacement for brawn," he said. "Every other time you've ever thought about technological change, you've probably been thinking about blue collar workers. This is a white collar revolution."
Wolfers said AI could bring about "once-in-a-generation" increases in worker productivity. He points to several case studies showing that the use of generative AI, particularly in large language models like ChatGPT, boosted productivity and quality of work in office-based workplaces. In one trial, staff were asked to craft an email to a coworker regarding poor performance at work.
"For those who were told to do it the way they normally would, it took 27 minutes to complete this task. For the folks who were given access to ChatGPT, it took them 16 minutes," Wolfers said. "That's a 40% rise in productivity. Additionally, the folks in the treatment group—those who used AI—not only did the same quantity of work, but higher quality work than their peers."
Despite these potential gains, AI has not been widely adopted in workplaces across the country.
"Only 10% of American businesses say that they've adopted AI. Another 14% suggest that they're going to do so within the next 6 months," said Wolfers. "We're not going to see [the effects of AI] in the productivity statistics yet, because work hasn't been redesigned."
Although AI represents positive changes for GDP and worker productivity, Wolfers warned the Appropriations Committee against allowing major AI companies to monopolize the market and capture most, if not all, profits that result from AI adoption.
"The problem is not the presence or absence of AI or of robots, but the market structure in the market for AI," he said. "Preserving competition is wildly important."
Watch: Wolfers' testimony to the Michigan House Appropriations Committee here.
