Wayne State teaches AI with real robots: industrial reskilling starts on the shop floor, not in the classroom

By Today@Wayne (Wayne State University) · November 10, 2026.

Wayne State University in Detroit is training its engineers with robots, conveyor belts, computer vision and generative AI at IntelliMake, a research center and pilot-scale factory it is developing with Kyndryl. According to the university, the program is already helping its graduates land jobs in the industry.

The backdrop is Michigan's state plan on AI and jobs. According to that plan, which the university cites, about 75% of manufacturing jobs will require some form of reskilling as AI and automation spread. The state also estimates that responsible AI adoption could generate up to $70 billion in economic impact and support 130,000 jobs. These are official projections, not measured facts.

The technical idea is simple. A traditional industrial robot follows instructions programmed down to the millimeter: it repeats tasks precisely, but fails if a part arrives slightly out of position or with a different geometry. In the lab, students experiment with systems that combine computer vision (cameras that interpret their surroundings) and AI agents (programs that decide and act) so the machine can adapt to the unexpected. The course IE 6000 "Digital Automation," redesigned by co-director Raj Sohmshetty, blends classroom and lab work.

The outcome data is modest and comes from the institution itself. Co-director Ratna Babu Chinnam says about 10 students worked intensively in the lab in fall 2025 and that several landed positions at prominent companies. One AI master's graduate, Gokula Kumaran Kiruthiga Gopal, was hired as an AI engineer at Panasonic and partly credits his hiring to his work at IntelliMake. These are testimonials, not an independent evaluation.

One figure is, in my view, more revealing than the rest: in a recent class, around 70% of Digital Automation students were already working engineers or professionals, including employees of automotive companies. An executive at PHINA (formerly Delphi Technologies) says he developed an AI agent that provides quick access to plant information and standard procedures, and that it eliminated several manual data-entry and administrative tasks.

Our take: the much-feared reskilling is already happening, and it is being done not through grand plans but by workers who go back to school while still employed. That fits a thesis we hold: AI doesn't simply "take jobs"; it changes which work matters. Administrative and repetitive work, as in the PHINA example, is the first to be absorbed by software. What gains value is domain judgment. Chinnam sums it up by saying computer scientists understand AI but don't always know what the real need is; that bridge between factory and algorithm is where the human advantage lies.

Still, it's worth not overselling. A lab with a dozen standout students and a handful of hires is a good sign, not proof that the transition is solved. In the short term, the 75% of jobs that will need reskilling implies an enormous effort, and not every worker will have access to a master's program or a nearby university. The risk of inequality between those who reskill and those who don't is real.

But the direction is right. A factory where machines adapt instead of breaking down, and where people supervise, design and improve rather than type in data, is a step toward the productive abundance we pursue in the long run. Testing ideas in a controlled environment, without halting a real plant, also lowers the cost of getting things wrong. If this model is replicated at more universities and opened up to those currently left out, the transition will be less painful than the numbers suggest.