Donald Trump says he’s bringing factories back to America. But here’s what he doesn’t say: that’s not the same as bringing back jobs.
Even with massive reinvestment, the jobs may never return. The “Golden Age” of American manufacturing won’t look anything like that of the 20th century.
Artificial Intelligence (AI) is driving a new wave of automation that’s transforming how economies work. And the hard truth is, we may bring the factories back, but the floors could remain empty. Here’s why.
US Manufacturing Hasn’t Slumped
People are often surprised to learn that U.S. factories are producing more goods than ever. The real difference is that they have far fewer workers.
A recent report from Brookings shows manufacturing output rising steadily from 2010 to 2016, with 10 to 20 percent increases in output but only a 2 to 5 percent increase in jobs. The slow growth in employment is largely due to advancements in automation and AI technologies.
In the late 1970s, nearly 20 million Americans worked in manufacturing—about a quarter of the workforce. Today, it’s under 13 million—less than 9%. So yes, the machines are humming.
But fewer and fewer workers are on the floor. A recent survey of U.S. manufacturing CEOs concludes that the real returns no longer come from new hires, but from AI.
Canada’s no different. The same shift is underway here: fewer hands, more systems.
What AI Is Actually Doing
At its peak, manufacturing made up a quarter of U.S. GDP. Today, it hovers around 11%—and falling.
Trump’s tariffs might raise output. But they won’t bring back the jobs of the past. The old assembly lines are gone.
Reshoring today isn’t about conveyor belts and forklifts. It’s about how AI is changing not just what gets made—but who makes it.
In smart factories, AI systems manage supply chains, run robotic arms, check products for flaws, and make decisions on the fly. Roles like assembly workers, foremen, and inspectors are disappearing or being rewritten. What’s replacing them? Technicians, data analysts, and AI supervisors.
Take Tesla. At several of its plants, AI-driven systems oversee everything from component assembly to quality control. The role of the human worker has shifted—from operating machinery to supervising algorithms. Factory jobs now require data fluency, constant adaptation, and systems thinking.
If the last automation wave was mechanical—machines repeating motions—this one is cognitive. These systems don’t just move. They decide. They adapt. They bring a kind of synthetic judgment that reshapes the very idea of work. And it’s not just factories. The same pattern is spilling into finance, logistics, customer service—even parts of the public sector. AI agents are already scheduling, analyzing, negotiating—even supervising.
This isn’t a glimpse of what’s next. It’s what’s happening now.
The Paradox: More Production, Fewer Jobs
Let’s name the paradox clearly because this is where Trump gets things very wrong. Tariffs may bring production back. But they won’t bring back jobs. And here are four reasons why:
- Tariffs don’t protect employment—they protect production. Modern factories are built to minimize labor and maximize output.
- Tariffs raise costs—but instead of hiring, companies invest in automation to stay competitive.
- The work that remains has changed. At Tesla, technicians manage systems, not materials.
- This isn’t a blip—it’s a shift. A new industrial logic is taking hold: fewer people, more systems, different skills.
Even if production returns, AI has already changed what that means. The industrial giants of the next decade aren’t expanding payrolls—they’re engineering them out.
SoftBank, a Japanese investment powerhouse, recently announced a $1 trillion strategy to fund a network of automated manufacturing hubs across North America. As for the jobs? “We are not building factories of the future—we are building the future without factories.” It’s a stark admission: AI will automate traditional roles out of existence.
The same for Foxconn, the Taiwanese electronics manufacturer best known for assembling Apple products. It was once hailed for reshoring jobs in Wisconsin. Today, it employs fewer than 250 workers at its plant. The rest of the floor space is filled with robotics. “The U.S. site will showcase our automation portfolio, not expand our workforce,” the company said.
Tesla is going even further—moving toward unmanned factories where, as Elon Musk puts it, “the goal is no lights, no workers—just bots, moving 24/7.”
These aren’t exceptions. They’re blueprints. And if governments want to prepare for the future of work, the real question isn’t whether factories are coming back. It’s who will be ready to work in them.
The Real Policy Challenge: Training and Transition
Trump’s reliance on tariffs isn’t new—but at a 145% on Chinese imports and 10% across other trading partners, the scale is. As for his goal of forcing production back to American soil, what Americans—and Canadians—really need isn’t tariffs. It’s a training revolution.
Workers must be ready for roles that are technical, adaptive, and AI-integrated. That means anticipating change—not resisting it—and designing policy for transformation, not nostalgia.
Amazon is a case in point. Despite criticism, the company is investing over $1.2 billion to retrain more than 300,000 workers by 2025, through programs like Career Choice and Machine Learning University. These efforts support transitions into cloud computing, robotics, and AI-powered roles.
If even automation giants like this are investing in people, it’s not hard to see where the real frontier lies.
The Populist Illusion
Trump’s economic nationalism doesn’t protect workers. On the contrary, it retools production for efficiency and capital—while telling voters it’s all for them.
This is an illusion. Trump may be bringing manufacturing home—but he’s also triggering a huge wave of automation. And no one’s making sure workers are invited.
AI poses real challenges for labor markets—and leaders should say so. Not with slogans about “better jobs,” but with policies that face the trade-offs, equip workers, and acknowledge that the machine-first economy is already here.
So why the silence? In our next column, we’ll ask what this avoidance says about Canada’s federal election—and why AI may be the most important issue no one wants to name.
Don Lenihan PhD is an expert in public engagement with a long-standing focus on how digital technologies are transforming societies, governments, and governance. This column appears weekly. To see earlier instalments in the series, click here.