A new US manufacturing boom may bring more AI than jobs

The United States is on the cusp of an automation boom in manufacturing.

Spurred in part by new tariffs on many imported goods, including 25% charges on foreign cars that President Donald Trump announced Wednesday, some companies have said they may shift select operations to America. New factories planned for the U.S. will not only be safer and more efficient. They will also be among the most modern in the world, helping America begin to catch up with manufacturing standouts like Germany and Japan.

Encouragingly, most companies are not aiming to eliminate humans from the factory floor. Instead, they hope to meld the functionality of robots with the adaptability of workers. That way, factories can quickly pivot no matter what happens: more tariffs, supply chain disruptions, or another unforeseen event.

To accomplish this, companies are leaning into artificial intelligence.

“The future of America is advanced manufacturing with AI at the center,” says Olaf Groth, a business and public policy professor at the University of California, Berkeley’s Haas School of Business and CEO of Cambrian Futures, an analysis and futurist firm for AI and emerging technology.

Propelled by tariffs

The push to set up automated factories in the U.S. stems in part from the tariffs that the Trump administration has imposed or threatened to impose. Companies can avoid adding to their costs by producing domestically. But the move to bring back American manufacturing started earlier, with the pandemic.

“All of a sudden, the brittleness of the [international] supply chains became laid bare,” says Chinedum Okwudire, professor of mechanical engineering and director of the Smart and Sustainable Automation Research Lab at the University of Michigan. Companies started shortening supply chains, bringing suppliers onshore, and automating factories to be more resilient in the face of uncertainty.

The change is already underway in some sectors. Spending on U.S. factory construction has tripled in the past four years, even after adjusting for inflation. Most of that has gone to already highly automated computer chip plants, spurred by subsidies enacted during the Biden administration. Now, it’s spreading to other sectors.

In its survey released in January, the Manufacturing Leadership Council found that a third of American manufacturers were scaling up AI-driven manufacturing while a slightly larger share was still just experimenting with it. A recent Deloitte survey said AI offered U.S. manufacturers one of the largest returns on investment of any technology.

The technology itself is racing ahead. Similar to how AI companies trained machines on hundreds of millions of pages of text so that ChatGPT can read and write text more convincingly, manufacturers are training computers on factory performance data and training robots on videos of factory work. That will allow machines to predict when robots will break down. It also makes those robots smarter and rapidly reprogrammable if the factory’s needs change. Another growing area: collaborative robots, called “cobots,” that work alongside human workers.

Eventually, robots could lead to big leaps in factory productivity. For example, they could be trained to make ever better versions of themselves.

“It will take time,” says Erik Brynjolfsson, director of Stanford University’s Digital Economy Lab. “But it’s happening faster than I expected, than most people expected.”

Other experts remain cautious. An automation boom could happen, says Dawn Tilbury, chair of the University of Michigan’s robotics department and former head of the National Science Foundation’s Engineering Directorate. But ChatGPT, as good as it is, still makes mistakes – and it’s getting the most robust training out there. With far less data and video to train factory robots than what’s on the internet to train ChatGPT, many companies are wary of using them in mission-critical areas. Downtime can cost a big factory $10,000 a minute, she points out.

How will humans fit in?

The biggest unknown involves how human labor fits in.

To optimists, the automated factory of the future will still hire low-skilled workers. “Manufacturing offers the greatest pathways for people to start in a low-skill position and work their way into high-skill,” says Dr. Groth of UC Berkeley. Automation also promises to make factory jobs simpler and, thus, more accessible to low-skilled workers, he adds.

Others are more pessimistic. The automation-related manufacturing boom may add slightly more jobs than it destroys, such as robot technicians, says Craig Le Clair, vice president and principal analyst at Forrester Research and author of “Random Acts of Automation.” Automation should boost production and jobs even if machines take over more jobs. “But it’s not going to be the significant labor ‘add’ that you had in the ’70s and ’80s.”

Editor's note: One sentence has been deleted so the story can end with the intended quotation from Mr. Le Clair.