Since 2000, the US has lost 5 million manufacturing jobs. Some say that the loss is due to foreign competition, primarily from China. Others say that it is due to higher productivity from automation and factories simply don't need as many workers. Historically, those two trends, globalization and automation, have reinforced each other.

Yet a study by Boston Consulting Group suggests that they are starting to diverge. As automation lowers production costs, the logic of offshoring isn't as compelling as it once was because transportation and other factors cancel out many of the benefits of moving to a low-cost country.

So we're going to rethink automation. We can no longer see it as a way to replace workers with machines that can work 24 hours a day without stopping for a coffee break, but as a way to increase productivity and create value by extending the work of humans. That means that if we are going to reap the full benefits of automation, we need to think more about people.

The New Age of Collaboration

There's no longer any doubt that machines can do work formerly done by humans, often not only more efficiently, but better. Does anybody really miss having to wait in line for a bank teller or having to call a travel agent to book a flight? Yet strangely, banks still have branches and travel agencies are still in business.

The reason is that, as Geoff Colvin points out in Humans Are Underrated, we increasingly need humans to do jobs that machines can't, chiefly working with other humans. He explains that just as the industrial revolution created the need for knowledge workers, this new cognitive era we're entering is creating a need for relationship workers.

Josh Sutton, who leads the data and AI practice at Publicis.Sapient, seems to agree. "I actually think cost is the worst reason to invest in automation, because the companies that are able to thrive 10 or 20 years from now will not be the ones who cut costs but those that transform their business models. You have to be looking to bring your organization up a level," he told me.

That the nature of work has changed has been obvious for some time in the professional class, where highly paid executives spend much of their time communicating with other highly paid professionals. But now, with the rise of collaborative robots, we're beginning to see the same trends arise in manufacturing too.

Aligning Production to Human Capital and to Customers

Traditionally, robots were designed to do rote tasks, such as painting a car body or welding parts together, much more efficiently than a human could. However, because they were unsafe for humans to be around, they had to work in fenced in cages, which meant that they required a lot of space. They were also expensive and difficult to retool to perform a new task.

However, the new collaborative robots, which can sense people and objects around them, easily work with humans. They are also much cheaper, with models selling for just over $20,000. That means you no longer need a large plant in a low cost region to be economically viable, but can use robots in smaller factories in urban areas that have an abundance of high skill labor.

That's important, because as Jeffrey Immelt, CEO of General Electric, explained in Harvard Business Review, engineers, designers and researchers need to work closely with production managers on the factory floor in order to continually innovate and improve quality. Another added benefit is that smaller production facilities can be located closer to customers.

Scott Eckert, CEO of Rethink Robotics, which makes the popular Baxter and Sawyer robots told me, "When factories are close to customers and smart robots enable small production runs, the supply chain becomes more agile as manufacturers can react more quickly to market and customer demands."

Empowering Workers Instead of Replacing Them

In talking with a number of robotics and artificial intelligence experts, probably the most compelling thing I found is that it's hard to find anyone who's talking about replacing workers. Just as you don't fire your lawyer because e-discovery software becomes available, the best way to make robots productive is for them to collaborate with humans.

"What we see among our customers is that they are using automation to increase productivity among their workers, rather than to reduce headcount" Eckert of Rethink Robotics says. "In fact, most of the companies we work with are experiencing a labor shortage, so they need to do everything they can to empower the employees that they have."

He also has noticed that the factory workers themselves have more to offer than anyone imagined. "It's often the people who were doing the job manually who have the best ideas about how to improve production processes. They are able to train the robots to optimize a process in ways you wouldn't expect. It's a remarkable thing to see.

It's not just the bosses who have become enamored with the robots either. "We have seen in many cases that not only does throughput improve significantly, but jobs are redesigned in a way that makes them more interesting and rewarding for the employee. Everybody wins," Eckert told me excitedly.

Sutton of Publicis Sapient agrees. "The best way to roll these technologies out and gain adoption is to create a great experience and that's the same internally and externally. You have to make your technology as compelling to your employees as you do to your customers. Automation, when properly understood, does not replace humans, but extends them."

Racing With The Machines

The Luddite movement arose in the early 19th century, because of the fear of tradesman that their skills, which had been handed down from father to son for generations, would soon be made obsolete by the industrial revolution. Wary of the future and consumed with fear and frustration, they smashed the machines in protest.

In a sense, we're all like the luddites now, because many of the skills that made us valuable in the past, such as the ability to retain information and apply it in the right context, are being uploaded to the cloud and programmed into robots. That's why, as MIT's Paul Autor points out, the jobs of the future are not white collar or blue collar, but those focused on non-routine tasks.

So the challenge today is much like in the early industrial revolution. Then, as now, there is little question that any task that can be automated will be automated, history has shown that to be the case. Rather, we need to think more clearly how automation can be used to empower workers so that they can better collaborate, with each other and with machines.

In other words, as Erik Brynjolfsson and Andrew McAfee explained in The Second Machine Age, the key is to not try to race against the machines, but to race with them. The key to winning in the new era of automation is not to eliminate people, but to put them at the center.