ATLANTA, United States — The United States Department of Defense is the world’s largest employer with almost 3 million personnel, including 1.3 million men and women on active duty. That’s a lot of people. It’s also a lot of uniforms.
But since 2002, a legal requirement known as the Berry Amendment has restricted the military from procuring clothing that is not made in the USA. Complying with the rule has proven challenging, however, for the same reason that 97 percent of all clothing purchased in the United States is imported: garment manufacturing is labour intensive and American workers are paid significantly more than their low-cost counterparts in Asia and elsewhere, making traditional production in the US prohibitively expensive.
In 2012, the military’s Defense Advanced Research Projects Agency (DARPA) — the emerging technologies unit credited with planting the seeds of the internet and spawning the development of self-driving cars — awarded a $1.26 million grant to a start-up called Softwear Automation, birthed thousands of miles from Silicon Valley on the university campus of Georgia Tech in Atlanta. Its mission: develop robotic sewing machines that could “produce garments with zero direct labour.”
In the garment manufacturing sector, people have worked alongside machines for hundreds of years. In fact, the original Luddites were English artisanal weavers who, in the early 19th century, rioted against the rise of the power loom, which automated a portion of their work. But the process of sewing finished garments has always required the dexterity of human hands — until now.
If robots are an efficiency strategy, it’s defending a dying model.
With a combination of machine vision, robotics and advanced computing, Softwear Automation has developed “sewbots” that can produce clothing without traditional garment workers, earning enough publicity to make the company a symbol of a new industrial revolution in garment manufacturing. “We created the term ‘sewbots,’ it is our trademark. There has always been automation historically, but sewing the garment together has always been done with seamstresses to guide the fabric. We have created a technology that replaces the need for a seamstress. For every four robots, you need one operator to manage,” explains Palaniswamy Rajan, CEO of Softwear Automation, which is targeting clients well beyond the Defense Department and has also raised money from The Walmart Foundation and CTW Venture Partners, a venture capital firm founded by Rajan himself.
At the moment, “sewbots” have yet to make complicated garments. “A fully completed, more complex garment being done by a robot is still some time away,” says Edward Hertzman, an apparel supply chain expert and founder of Sourcing Journal. “I don’t think it’s going to come to mass scale in the near future.”
For now, Rajan is focused on exploiting the near-term opportunity in simple high-volume garments like t-shirts and eventually denim. “It’s like a baby learning more skills and becoming more capable and dexterous,” he says of the technology. “You also need a certain volume of product to be economically viable. Right now, if you make 1 million pieces of something, then ‘sewbots’ are great for you. The businesses that are doing t-shirts are big. This is our focus for next year.”
Rajan sees a future where demand for clothing will grow enormously as cheap labour becomes scarce, making automation essential. “I’ll give you the macro points: people are living longer and you can expect to have population growth of 3 billion more people worldwide who all need food, shelter and clothing,” he explains. “Meanwhile, the age of cheap labour is going to end; there will be less places with cheap labour to exploit. Today, the average age of a seamstress in the US is 56. In 10 years, you won’t have any. It’s a line of work that’s not attracting people. In China, you have worker shortages and rising wages. In India, you have the same thing: young people not wanting these jobs.” Even in places like Cambodia, Laos and Bangladesh, wages are rising as labour shortages grow.
“Quality is also important,” he continues. “People make mistakes. We had a customer who had a 2 percent error rate in their production; our robots operate at a 0.7 percent error rate.” But as Rajan agrees, the opportunity in automation extends far beyond consistent quality and the cost efficiencies of replacing human labour.
The mismatch right now is that a digital consumer is being served by an analogue supply chain.
The rise of “sewbots” challenges the very foundations of the current mass fashion model. “I think the impact will be incremental and modest if the goal of robotics is to lower costs through fewer workers. That is an argument for efficiency, but efficiency alone is not really strategy. Making the current industry model — volume, low-cost, seasonal and Asia-based — more efficient will not really transform performance or profitability,” says John S. Thorbeck, chairman of Chainge Capital, a supply chain advisory firm.
Leading Chinese apparel companies, like Esquel and TAL, have been investing in automation technology for years. “So while costs are rising, their investment in productivity through automation is keeping them competitive, but ultimately I think that’s defensive,” explains Thorbeck. “Wages are not the highest cost to consider; the cost of labour is actually diminishing as a percentage of total cost. The highest cost is clearly the cost of mark downs and lost sales.”
Thorbeck sees the ultimate power of automation, not in reducing labour costs, but in boosting overall productivity by bringing greater speed and flexibility to supply chains, enabling retailers to be more responsive to consumer demand in a world where expectations and trends are increasingly shaped at the speed of digital media. “The mismatch right now is that a digital consumer is being served by an analogue supply chain. It’s really about aligning information flows, bringing a manufacturer closer to demand visibility. This goes hand in hand with automation,” he says. “If robots are an efficiency strategy, it’s defending a dying model. The new model is shorter production runs with the idea of reducing risk, but not necessarily cost. It’s about lead time optimisation.”
Not only are robots faster and more flexible than human workers, they also free retailers to move some elements of production to Europe and the US, closer to both product designers and large centres of demand, creating an even tighter loop with consumers. “I think there’s an opportunity for more localised production and that’s the idea athletic companies are beginning to explore,” notes Thorbeck.
German sportswear giant Adidas has launched a new automated “Speedfactory” in the Bavarian town of Ansbach, where the company is deploying new processes like robotic cutting, computerised knitting and 3D printing. Adidas is also building a second “Speedfactory” near Atlanta, Georgia, targeting the American market.
“It’s a hyper- flexible shoe factory concept that can be placed anywhere in the world. It enables us to combine speed in manufacturing with the flexibility to rethink conventional processes to give consumers what they want when they want it,” explains Gerd Manz, the company’s head of technology innovation. “It allows us to react quickly to consumer demands and changing trends. The vision is to reduce lead times from months to weeks to days or hours. Consumers today live in a constantly changing world. This shapes their behaviour and expectations. They demand newness and immediacy without compromise.”
The new model is shorter production runs with the idea of reducing risk, but not necessarily cost. It’s about lead time optimisation.
American sportswear behemoth Nike is also experimenting with automated manufacturing techniques. “We’re expanding into more styles with advanced automation in the build and adhesion processes, localising sourcing for quicker material lead time and reducing waste through advanced knitting and higher-precision cutting technologies,” said Mark Parker, the company’s chairman, president and CEO, on a recent earnings call.
“We’re lining out the pipes of our Manufacturing Revolution with real-time consumer insights, seamless inventory and a hyper-local marketplace approach, all of which is connecting the dots in the name of speed for the consumer.”
Inditex, the parent company of Zara, which operates a number of automated factories in Spain, is said to be increasing its focus on robotic manufacturing, as is Fast Retailing, Uniqlo’s parent company, though both firms declined to comment.
But as fashion, a major global employer, embraces robotic automation, what are the implications for millions of garment workers around the world?
While new technologies can certainly displace large numbers of workers, they also boost overall productivity and create new higher value, higher paid jobs. “The difference between today’s life expectancy and material living standards in rich countries compared to 100 years ago is almost completely about automation, which took people out of doing things in traditional ways and created more output with less labour input, which is higher productivity, and freed up people to do other higher value-added activities,” explains Lawrence Katz, an economics professor at Harvard who studies labour and technological change, and served as chief economist at the US Department of Labor when Bill Clinton was president.
Take the classic example of the automobile. When Ford Motor Company began producing its Model T, horses became obsolete and blacksmiths, stable boys and carriage drivers found themselves out of work. But of course, cars also created new higher value, higher paid jobs, like auto mechanics. They also gave rise to new industries like road construction, motels and fast food, raising net employment.
“In apparel, there have been major historical innovations that have reduced the labour input to different parts of the industry and employment has plummeted,” says Katz. “But the next generation got improved education, there were opportunities in other sectors and the transition went reasonably smoothly — although particular areas and individuals were affected. Look at the old textiles towns in northern England. You can see that sudden shocks clearly have impact. All new technologies create winners and losers.”
In the garment factories of Southeast Asia, up to 90 percent of workers could lose their jobs due to the rise of “sewbots” and other disruptive technologies like 3D printing, according to a 2016 report by the International Labour Organisation (ILO) on how technology is changing labour markets in the ASEAN region, which includes important apparel exporters like Cambodia and Vietnam, but not China, India or Bangladesh. “The region will encounter both a displacement of lower-skilled workers and an increase in the demand for higher-skilled technicians and engineers,” explains the report. But what will the net effect be?
One issue will be how the benefits of robotic automation are distributed. A small group will own access to the technology.
At present, it’s difficult to predict. “Automated sewing machines really being able to create finished three-dimensional products — that might be another two to three years away,” says Jae-Hee Chang, co-author of the ILO report. “But there are 9 million sewers in the ASEAN region alone. If going forward, you just need hundreds of thousands of people — you know, with more skills and better pay— that means the remaining workers will need to find opportunities outside the sector,” she continues. “But if there’s no other sector that can absorb them, that will pose a problem. People in Cambodia who we spoke to basically said: ‘If there are no jobs for garment workers, there’s going to be another civil unrest in the country.’”
“I believe automation is good, as people aren’t meant to sit there and behave like machines. But I am concerned about places like Bangladesh that are dependent on this work,” said Jonathan Zornow, whose fledgling, Seattle-based start-up Sewbo is also targeting the “sewbots” opportunity, using a solution that is much simpler than its competitors. (Rather than develop complex robots with the dexterity to sew clothing, Zornow’s approach uses water-soluble thermo- plastics to temporarily stiffen fabrics, which can then be manipulated by existing industrial robots).
“Preparing for change is the key,” says Katz. “One issue will be how the benefits of robotic automation are distributed. A small group will own access to the technology and we could have rapidly rising total output but a large group of people that are disconnected from the benefits. In worst-case scenarios, it could lead to divided societies and hostilities. But from the mid-20th century to the 1980s, we saw a shared prosperity event, though there was just as much new automation; at certain junctures we have simply done a better job of dealing with it. Maybe automation will be different this time, we don’t know.”
Many believe this time is indeed different. For one, technology is advancing faster than ever before and yet the economic benefits have not been widely distributed. What’s more, the rise of artificial intelligence-assisted automation, while productivity boosting, could eliminate a much wider range of jobs—not just low- skilled jobs, but high-skilled jobs that involve a large number of repeatable tasks.
Some have suggested that, as jobs disappear, governments will have no choice but to pay their citizens a universal basic income. But the idea faces deep cultural obstacles. “You want a safety net direct income,” says Katz. “But that does not solve the problem of providing meaningful work.” But do people need work? Or could a radically automated, post-work world liberate humans from daily toil?
“Work is a way of organising life that brings respect and meaning in a way that just an income transfer does not do,” says Katz. “When you look at parts of the US and UK where work has disappeared and where there is a safety net, the problem is not material wellbeing. Guaranteed income won’t be the whole solution. It’s playing defence, not offence... It’s about training the next generation for a new economy.”
Microsoft founder Bill Gates recently suggested that robots should be taxed to slow the pace of automation and fund programs to financially support and retrain displaced workers. But the proposal has met with widespread criticism.
“I don’t think directly targeting robots makes sense,” says Katz. “I think a general income or wealth tax is a much more sensible and less distortionary policy. We don’t want to tax innovators.
“Robots are not the enemy.”
This article appears in BoF's special print edition: "America."
Did you know that BoF Professionals receive our print issues first? Annual BoF Professional memberships also include unlimited access to articles, exclusive analysis, invitations to networking events and the members-only app. Not a BoF Professional? Subscribe here. The issue is also available for purchase at shop.businessoffashion.com and at select retailers around the world.