The impact of sewing robots on emerging countries
In a large factory workshop, Kiran Kumari, dressed in purple and gold saris, sits in front of a Brother sewing machine, surrounded by dozens of garment workers colleagues. With a skillful wave of her arm, she picked up a piece of cloth from the bundle of cloth next to her, put a white patch on it, and quickly operated it under the needle. This is only one of about 400 collars sewn by Kumari for Polo Ralph Lauren's jacket during the 8-hour working hours. It takes a few minutes to complete a collar, and Kumari’s monthly salary is about 100 dollars.
She and about 4,800 other garment workers are scattered in three Matrix garment factories south of Delhi. They are a small part of the cheap labor force that many developing countries hope to promote the country to prosperity. This is especially true for South Asia, where the population is growing rapidly and wages are still low. The World Bank estimates that during the next 20 years, 1 million to 1.2 million new workers will enter the labor market every month in the region alone, adding a total of 240 million laborers.
However, in Atlanta, 8,000 miles away, a robotics company is working to develop a machine that may leave Kumari out of work forever. The Sewbot robot technology developed by Softwear Automation is designed to automate the entire garment making process.
The technology still needs many years to achieve low cost and reliability enough to replace manpower. For example, Kumari earns about $1,200 per year. The company will not say how much the Sewbot robot will cost, but industry sources say the price is as high as hundreds of thousands of dollars. But as automation sweeps across mature industries, experts warn that it’s only a matter of time before this technology disrupts a large part of the economic model of the developing world.
Considering the extent to which South Asia’s economic plans depend on undertaking international manufacturing operations (China has become too expensive for this type of business), the region faces exceptional risks. Policymakers in India, Pakistan, and Bangladesh are discussing reaping the “demographic dividend” because the populations of these countries are growing rapidly, and the average wage is still only about one-fourth of the level in China.
However, economists are beginning to ask how much dividends can these young, cheap, and potentially unstable laborers enjoy as robots gradually become capable of performing the heavy manual labor they depend on for a living? Southeast Asia (another clothing manufacturing hub) and other regions such as sub-Saharan Africa may also feel the impact.
"Robots and artificial intelligence are the next industrial revolution," said Rajiv Kumar, an economist and founder of the Pahle India Foundation. , Electricity, assembly lines or computers-more disruptive, because they will not only replace routines, but also complex brain functions. What is worrying is that our so-called demographic dividend will become a demographic nightmare."
A few years ago, scholars noticed that some strange things were happening in parts of Asia, Latin America, and sub-Saharan Africa. Although many countries have been experiencing strong growth, the share of manufacturing jobs has hardly increased since the 1980s, and has even begun to decline in some countries-much earlier than economists expected.
In 2015, Harvard economist Dani Rodrik proposed the concept of "premature deindustrialization." He observed that compared with the development process of Western countries, many developing countries have turned to a service-led economic model at a much earlier stage of development. He said that technological change has played a big role in this, and warned that this trend may have serious consequences for economic growth and political stability in these regions. “Manufacturing has traditionally absorbed a large amount of unskilled labor,” he wrote, and he also warned that the trend of deindustrialization “is not necessarily a good thing for liberal democracy.
Roderick’s findings help explain a study conducted by the International Labour Organization (ILO) and the United Nations Development Programme (UN Development Programme) shortly after the financial crisis. The report found that between 2003 and 2009, South Asia's employment growth rate was only one-third of the overall economic growth rate. Experts warn that the region is experiencing "no employment growth."
Since then, the pace of technological change has accelerated, and some industries have even lost jobs. Indian IT service companies that have prospered in the past 30 years have begun to give way to automated cloud computing systems. The two largest companies in the industry, Infosys and Tata Consultancy Services, both laid off staff this year. At the same time, at the Hyundai car plant in Chennai, 400 robots have replaced humans in a large part of the production process.
image.png
The chief executive of a large Indian technology company (he asked not to be named) said that if the bosses were not worried about the consequences of massive layoffs, the layoffs would be even greater. "We conducted an audit and found that we can replace half of our employees with artificial intelligence," he said.
"If it were not for the social shock wave of doing this, we would really do it."
The clothing industry is the dominant industry in these economies. Bangladesh is particularly dependent on the clothing industry, which accounts for 82% of the country’s exports, and 2.5% of the country’s population is engaged in clothing manufacturing. Statistics from the Clean Clothes Campaign, which strives for better working conditions for garment workers, show that in Bangladesh, India and Pakistan, a total of about 27 million people are employed in the apparel industry.
One reason the apparel industry is creating jobs is that workers in South Asia are competing at lower wages than their Chinese counterparts. According to data from market research firm Euromonitor, the average hourly wage of Chinese factory workers reached 3.6 US dollars last year, nearly four times that of 10 years ago. The average salary of Chinese factory workers is now about five times that of their Indian counterparts, which is close to the level of Portugal or South Africa.
Another reason is that, surprisingly, the apparel industry has proven to be less susceptible to automation. Since the advent of automatic sewing machines in the 19th century, the process of making a T-shirt has hardly changed. T-shirts are usually made of cotton, which is a thin and light material that curls easily, making it difficult for robots to pick up and move accurately. Behind Kumari's handy quick stitching behavior is a series of tiny processes and decisions, which are so complicated that even the most competent software engineers can hardly replicate them.
"There are four processes to make a garment," said Gautam Nair, the general manager of Matrix Clothing. "Pick up the fabric, align, sew and set it aside. Of these processes, only sewing is currently achieved. With automation, sewing machines appeared a long time ago. It’s still faster and cheaper to complete the other parts of the production process manually."
This situation may be about to change. In an old factory in Atlanta, a group of technicians from Softwear Automation typed computer code on the touch screen to try to perfect the Sewbot robot. In order to solve the problem of sewing delicate fabrics, they installed a camera that acts as an eye for this robot, just like what engineers did for self-driving cars. These cameras can take pictures of the fabric being stitched, analyze the images, and guide the movement of the robotic arm accordingly.
This technology caught the attention of Walmart, the world's largest retailer, which gave the company $2 million as part of an automated jeans production project. In September last year, the Sewbot robot achieved a breakthrough, successfully stitching an outer stitch to a pair of jeans.
Next year, the company plans to use the Sewbot robot to produce T-shirts, and currently 97% of T-shirts are produced outside the United States.
Palaniswamy Rajan, CEO of Softwear Automation, explained the business necessity behind the development of such technologies, even if clothing companies have a lot of cheap labor to use.
"If you are Macy's and want 100,000 pieces of this style of clothing, you have to order from China 9 to 12 months in advance," he said. "If you make the place of production closer to the consumer, you can The next order of 10,000 pieces will be delivered within one month, and then we can see if some designs will sell better than others."
Wal-Mart is not the only major retailer focusing on automated garment making. In April of this year, Amazon applied for a patent to develop a "sewing on demand" machine that can automatically produce clothing after receiving an order. However, it is not only rapid technological change but also political reality that stimulates American companies. The election of Donald Trump as President of the United States and his “America First” trade policy commitments have prompted companies to consider how to move jobs back to the United States.
"There is no doubt that our work will be very different in the future," Walmart CEO Doug McMillon recently told employees. "Robots, drones, and algorithms will do some of the work we had to do in the past. Some people Fear of the results of these changes. I don’t think we need to be afraid... the secret of our success will always be our employees."
Policy makers in developing countries are also aware of these trends. Some of them believe that as long as the government has a good grasp of policies, there is still time to create a manufacturing boom in South Asia.
Arvind Subramanian, chief economic adviser to Indian Prime Minister Narendra Modi, has studied the issue of "premature deindustrialization". Sitting in a large office in the Ministry of Finance building in New Delhi, he said: “Yes, robots have already started cutting soft fabrics. But I don’t know if this is our next 9 or 10 years-or more realistically, The next 20 years-something to worry about."
Subramaniya said that combining good training, well-arranged free trade agreements and labor market reforms can create manufacturing jobs before robots become mature and cheap enough to replace workers. "If he is wrong, the social impact will be huge," warned Kumar of the Pahle India Foundation. "This may be a nightmare, because all educated and aspiring young people will be affected by this automation. And unemployed."
Others believe that these countries should take "stop loss" actions and focus instead on the service industry. In 2015, Uri Dadush, a senior researcher at the OCP Policy Centre, a Moroccan think tank, wrote that “the importance of manufacturing as a development lever is declining”. He believes that countries such as India, Pakistan and Bangladesh should focus on transforming themselves into specialized centers for certain industries such as tourism, transportation or finance.
For Kumari, the prospect that the service industry employment boom will make up for the loss of manufacturing employment has no comforting effect.
"If I don't make clothes, I have nothing else to do," she said. "I haven't received any other training."
How robots are better at making clothes than humans
Seattle software developer Jonathan Zornow tinkered with a method last year-he believes this solution removes the technical barriers that have restricted sewing automation for decades. At the time of "extremely boring", the 28-year-old programmer watched Discovery Channel's documentary "How It's Made" in one breath, during which he was surprised to find that he completed 38 steps of making a pair of jeans. At least 9 people are required for the process.
Zono’s approach is different from others: he wants to change the fabric structure to make the fabric better match the robot's operation, rather than improve the robot's ability to handle the fabric. He applied for a patent for a new process he developed: soaking the cloth in a melt of thermoplastic plastic-making cotton and other delicate materials as hard as wood. The robot then sews, stitches and shapes the hard fabric. Once washed with hot water, the fabric will regain its elasticity and become a pair of jeans. Zono said that he used this method to make the first clothes made entirely by robots: a T-shirt.
Zono said he is in talks with "some very large retailers" and manufacturers in China, India and Sri Lanka to promote the technology called Sewbo. Last month, executives from Adidas and Eileen Fisher, an American fashion design company, watched a Sewbo demo in Manhattan.
But even fanatics like Zono admit that, at least for now, cheap labor has a competitive advantage over machines (such as his robots). He said, "If you are in Bangladesh, when you have a house and people who are willing to work for you for a salary of $1 a day, you will not buy a super advanced sewing machine."
Instead, he sees sewing robots as an opportunity to bring manufacturing closer to shoppers and shorten the long and sluggish supply chain-each T-shirt takes an average of about 20,000 miles to reach the customer.