Blue-collar jobs will survive the rise of artificial intelligence. But the work will change

It’s hiring day at Rolls Royce’s jet-engine plant near Petersburg, Va. Twelve candidates are divided into three teams and given the task of assembling a box. Twelve Rolls Royce employees stand around them, one assigned to each candidate, taking notes.

The box is a prop, and the test has nothing to do with programming or repairing the robots that make engine parts here. It’s about collaborative problem solving.

“We are looking at what they say, we are looking at what they do, we are looking at the body language of how they are interacting,” says Lorin Sodell, the plant manager.

For all the technical marvels inside this fully automated, 8-year-old facility, Sodell talks a lot about soft skills such as trouble shooting and intuition.

“There are virtually no manual operations here anymore,” he says. People “aren’t as tied to the equipment as they were in the past, and they are really freed up to work on more higher-order activities.”

Call it the automation paradox: The infusion of artificial intelligence, robotics and big data into the workplace is elevating the demand for people’s ingenuity, to reinvent a process or rapidly solve problems in an emergency.

The new blue-collar labor force will need four “distinctively more human” core competencies for advanced production: complex reasoning, social and emotional intelligence, creativity and certain forms of sensory perception, according to Jim Wilson, a managing director at Accenture Plc.

“Work in a certain sense, and globally in manufacturing, is becoming more human and less robotic,” says Wilson, who helped lead an Accenture study on emerging technologies and employment needs covering 14,000 companies in 14 large, industrialized nations.

Few narratives in economics and social policy are as alarmist as the penetration of automation and artificial intelligence into the workplace, especially in manufacturing.

Economists talk about the hollowing-out of middle-income employment. American political discourse is full of nostalgia for high-paying blue-collar jobs. The Trump administration is imposing tariffs and rewriting trade agreements to entice companies to keep plants in the U.S. or even bring them back.

The stark reality is that automation will continue to erode repetitive work no matter where people do it. But there is also a myth in this narrative that suggests America has permanently lost its edge. The vacant mills in the southeast and Midwest, and the struggling cities around them, are evidence of how technology and low-cost labor can rapidly kill off less-agile industries. This isn’t necessarily a prologue to what’s next, however.

Cutting-edge manufacturing not only involves the extreme precision of a Rolls Royce turbofan disc. It’s also moving toward mass customization and what Erica Fuchs calls “parts consolidation” — making more-complex blocks of components so a car, for example, has far fewer parts. This new frontier often involves experimentation, with engineers learning through frequent contact with production staff, requiring workers to make new kinds of contributions.

“This is a chance for the U.S. to lead. We have the knowledge and skills,” says Fuchs, an engineering and public-policy professor at Carnegie Mellon University. “When you move manufacturing overseas, it can become unprofitable to produce with the most advanced technologies.”

The new alliance between labor and smart machines is apparent on Rolls Royce’s shop floor. The 33 machinists aren’t repeating one single operation but are responsible for the flow of fan-disc and turbine-blade production. They are in charge of their day, monitoring operations, consulting with engineers and maintaining equipment.

This demonstrates what automation really does: It changes the way people use their time. A visit to the plant also reveals why factory workers in automated operations need more than some knowledge of machine-tool maintenance and programming: They are part of a process run by a team.

Sodell opens what looks like a giant suitcase. Inside is a titanium disc about the size of a truck tire. Unfinished, it costs $35,000, and it’s worth more than twice that much once it’s machined as closely as possible to the engineers’ perfect mathematical description of the part. The end product is so finely cut and grooved it resembles a piece of industrial jewelry.

“I am not at all bothered by the fact that there isn’t a person here looking after this,” he says, standing next to a cutting station about half the size of a subway car. Inside, a robot arm is measuring by itself, picking out its own tools and recording data along the way.

Variations in the material, temperatures and vibration can cause the robot to deviate from the engineers’ model. So human instinct and know-how are required to devise new techniques that reduce the variance. Just by looking at the way titanium is flecking off a disc in the cutting cell, for example, a machinist can tell something is off, Sodell says. With expensive raw materials, such technical acumen is crucial.

It’s also important because current artificial-intelligence systems don’t have full comprehension of non-standard events, the way a GPS in a car can’t comprehend a sudden detour. And they don’t always have the ability to come up with innovations that improve the process.

Sodell says workers are constantly looking for ways to refine automation. He tells the story of a new hire who figured out a way to get one of the machines to clean itself. He developed a tool and wrote a program that is now part of the production system.

Technicians start off making $48,000 a year and can earn as much as $70,000, depending on achievement and skill level. Most need at least two years of experience or precision-machining certification from a community college.

Rolls Royce is collaborating with these schools and relying on instructors like Tim Robertson, among the first 50 people it hired in Virginia. He now teaches advanced manufacturing at Danville Community College and says it’s hard to explain what work is like at an automated facility. Jobs require a lot more mental engagement, he explains, because machinists are looking at data as much as materials and equipment.

The Danville program includes a class on talking through conflict, along with live production where students are required to meet a schedule for different components in a simulated plant. The group stops twice a day and discusses how to optimize work flow.

“You can ship a machine tool to any country in the world,” Robertson says. “But the key is going to be the high-level technician that can interact with the data at high-level activity and be flexible.”