Minnesota workers at risk with highest generative AI ‘exposure’ in Midwest

Minnesota’s workforce is more exposed to generative AI than any other in the Midwest, according to a recent report from North Star Policy Action and the University of St. Thomas. 

The study found that 812,690 Minnesotans, or 31.3% of the state’s workforce, are in jobs where generative AI could partially or fully handle at least half of their tasks. Another 110,460 workers are in “very highly exposed” roles, where more than 75% of tasks could be affected. That puts Minnesota first in the Midwest and 10th nationally.

In the list of Midwest states measured, Michigan was next ‘exposed’ with 30.4% of workers, followed by Wisconsin with 29.3%, Illinois with 28.4%, Ohio at 28.1% and Indiana with ‘only’ 26.1%

‍

‍

That is a more serious warning than the group’s January 2025 estimate, which put the figure at about 17% of the workforce, or roughly 500,000 workers. The earlier analysis had used a Methodology that was implemented before ChatGPT was released. The new report uses a GenAI-specific framework and makes a point worth noting: exposure is not the same thing as replacement. In the report’s definition, “high exposure” means AI could handle half or more of a job’s tasks. The report emphasizes that it could mean layoffs, or that jobs are reworked, sped up or more tightly managed rather than simply eliminated outright.

The bigger story is where the report believes this exposure sits, in that it is not primarily on the factory floor as many commentators have suggested. The largest high-exposure occupations in Minnesota include customer service representatives, office clerks, software developers, manufacturing sales representatives, accountants and auditors, secretaries and administrative assistants, and bookkeeping clerks. Customer service representatives account for the largest number of workers, at 59,170. Office clerks follow at 55,710. Software developers score especially high on exposure, at 87.

At the industry level, finance stands out. The report says 79% of finance workers in Minnesota are in highly exposed jobs. Management follows at 75.7%, then professional services at 66% and information at 61.4%. Even sectors that sound less obviously vulnerable still show meaningful exposure: 33.7% of manufacturing workers and 28.3% of education workers are in highly exposed roles. In plain terms, generative AI is not just coming for coders or copywriters. It reaches deep into routine, document-heavy, analysis-heavy and coordination-heavy work across the state economy.

The exposure is not evenly distributed. The report finds that nearly 40% of women in Minnesota are in highly exposed jobs, compared with 34% of men. It also finds that exposure tends to be higher in occupations and industries with lower union density. That may potentially leave many of the workers most likely to have their jobs at minimum reshaped by AI with less bargaining power when such technology is deployed.

That may also help explain why the issue has drawn sharp interest in the Minnesota state Capitol, where local representatives have made AI regulation a priority. The report strengthens the argument for such regulation by saying the real reach of AI is beyond task automation. It points to electronic monitoring, algorithmic hiring and firing decisions, and other forms of AI-backed workplace control that may impact workers well before a job disappears. It also argues that Minnesota’s labor-related AI legislation has remained stagnant even as adoption has accelerated.

University of Minnesota’s Galin Jones, the school’s inaugural vice provost for AI, told the Minnesota Daily that institutions still have not worked out what useful, field-specific AI training should look like, but that “people who are prepared to use it in the workforce have more opportunity,”

The question is no longer whether generative AI will touch only Minnesota, or also the wider Midwest’s white-collar economy. It already has. The more consequential question is probably how the future will get shaped and who gets to define the terms.

‍