The field of robotics is hot right now. So is “disembodied” AI, but where we have already seen models like ChatGPT roll out to consumers, we have not seen household robotics do the same. As for business robotics, we’re still dealing with a profound mismatch where human workers are worried about being displaced.
But what if we could solve two of our greatest problems, killing two big birds with one rapidly thrown stone, if you’ll pardon the metaphor?
One of those problems is the death of manufacturing in America. This article chronicles the decline happening between 2000 and 2010, with so many jobs lost, and the quality of manufacturing jobs in America (i.e. the pay) worsening. In this telling, MIT political scientist Suzanne Berger talks about the threat of cheap Chinese labor.
“Jobs were killed by [China’s cheaper manufacturing goods],” she said. “Their cheaper imports: that struggle won the election for Trump.”
But the equation has changed: now Chinese labor is more expensive. At the same time, we’re not seeing the jobs return. There’s a broad consensus that this has to do with both the cost of labor, and a skills gap. That leads us to the question: what if we could use robotics to boost American manufacturing, with American jobs for humans at the helm, jobs that pay a living wage and are inherently skilled?
It sounds like a win-win.
Let the Robots Work
Some economists and analysts are already making this point, noting the ineffectiveness of tariff activity and acknowledging the reality of traditional labor costs.
“It’s become increasingly clear that the only cost-effective way to bring manufacturing back to U.S. soil is to automate unskilled labor via robotics while human operators, technicians, and managers oversee complex production systems,” writes Tom Biegala for Fortune. “This is not labor replacement, but a state where humans direct and optimize robotic automation. While the path to building hundreds of automation lines in the U.S. is not simple, the trade conflict puts manufacturing investment at an inflection point. After clearing certain technological and manufacturing hurdles, we can seize the opportunity to onshore by utilizing today’s class of AI-enabled robotic solutions.”
Sohrab Haghighat says it a slightly different way. In a recent presentation at Stanford, the CEO of Hestus, a CAD software business, talked about the opportunity we have to reframe American manufacturing.
“We have over half a million unfulfilled manufacturing jobs in the U.S. today,” he said. “That is the kind of work that you cannot do remotely. That is the kind of work that requires deep hands-on experience and skill, and it’s hard work.”
He contrasted this with the time when layoffs signified the demise of American manufacturing jobs:
“It became even harder to make a decent living in manufacturing, so fewer and fewer people wanted to do it,” he said. “But here’s the thing, the same autonomy that got us into this trouble can take us out of it assuming that we understand every problem is not a nail and not every robot has to be a humanoid. (We) just had a humanoid on stage, half an hour or an hour ago now, that serves certain applications, but not every application is served best by a human.”
So it’s necessary to look at workloads on a task-by-task basis.
The Glitch in Engineering for Manufacturing Systems
In addition, Haghighat laid out another issue with innovation stateside.
“Every product is designed on a computer model,” he said, “And engineers who do that design know nothing about manufacturing and manufacturing constraints. They only learn about it when they finish the design, send it to the manufacturer, and receive feedback. As a result, they spend more than 50% of their time redoing the work that they have already done. That’s $400 billion of lost productivity every year.”
With that in mind, he posited a scenario where every human engineer works with an AI agent that understands the needs and how manufacturing networking happens.
“You’re pairing every engineer with an AI manufacturing expert,” he said, “So as they go through the design, they receive the same feedback that the experienced manufacturer would have provided to them in real time. So by the time they get to the end of the design, it’s ready to manufacture.”
Here’s how he concluded by re-stating the proposition:
“It takes far more than incentives and tariffs and whatnot to bring manufacturing back to the U.S.,” he said, echoing the above view on tariffs as insufficient. “And when we talk about bringing manufacturing back to the U.S., we’re not talking about time-traveling back to the ‘60s. We need to forge a new reality. Here, the knowledge of manufacturing used to be passed from a person, to a person, to a person, and we have broken that chain. And we can whine about it, or we can do something different. And AI is the tool that has infinite capacity of absorbing human knowledge, and understanding and offering that as a tool to every engineer who is designing a product, so we can completely get rid of that design for ‘manufacturing that comes after manufacturing.’”
Think about this playbook as we ponder returning manufacturing to domestic shores. Already, smaller shops are starting to adopt this strategy. We may soon have a stratified system where the humans run the machines, for great labor savings, and everything that comes with that, without the high unemployment that people are afraid of. It just means being smarter.
