KERICHO, Kenya, and SAVAR, Bangladesh—Rose Nyunja was just 18 when she began working in the tea plantations of Kericho, Kenya’s biggest tea-growing region and a major source of employment for poor women in the country. For decades, she toiled away in the tea gardens, picking the leaves by hand.
Then came the harvesting machinery. Women like Nyunja started to lose their jobs by the thousands to machines that could each replace more than 100 workers.
One evening in 2020, Nyunja returned to the staff quarters to find her front door barricaded. She’d been fired. Nyunja pleaded with her supervisor to save her job—and her home. Instead, company security ejected her from the compound.
“My 26 years of service meant nothing to them,” she said, fighting back tears. “I was given one hour to remove my household items and leave. I have never experienced such humiliation and embarrassment in my life. I worked diligently for over two decades, and what have I got? Nothing.”
As Kenya’s tea estates automate to improve productivity, workers like Nyunja represent a broader global trend: Women are more likely than men to lose their jobs. A 2019 study by Britain’s Office for National Statistics found that 70 percent of jobs at high risk from automation are held by women. This April, a University of North Carolina study found that almost 80 percent of women in the U.S. workforce will be affected by advances in generative artificial intelligence, compared with 58 percent of men.
While recent advances in generative AI have sharpened concerns about loss of jobs for white-collar workers, job losses due to companies ramping up automation have been taking place for years, as seen in Kenya. Kweilin Ellingrud, a director at the McKinsey Global Institute, said her research shows that automation is 14 times more likely to impact low-wage workers than high earners.
“I think the reason it’s grabbing headlines is because it is also affecting higher-wage jobs for the first time,” Ellingrud said. “I think now generative AI is focusing and impacting jobs across the spectrum—it affects your job, it affects my job. Some of us, myself included, aren’t used to thinking: ‘How will my work have to change? How will my job change?’”
A solar-powered robot gathers tea sprouts using automatic optical identification in a tea garden in Hangzhou, China, on March 28.Long Wei/Future Publishing via Getty Images
In Kericho, Roselyne Wasike, a tea picker, said her income has fallen by almost half, from $150 a month to $80 a month, as the machines take over larger shares of work on the plantation. Even those who have managed to keep their jobs cannot escape the impact of automation. Many of them are widows and single mothers.
“These machines have disadvantaged women by making them redundant instead of giving them other tasks within the tea estates,” Wasike said.
The Kenya Plantation and Agricultural Workers Union said 30,000 women have lost their jobs as a result of automation in the past five years. About 60 percent of the 75,000 workers currently employed in the tea sector are women, down from an estimated 75 percent in 2017, according to Dickson Sang, the secretary-general of the union.
The resentment toward the machines boiled over into violence this May. Kericho residents torched nine harvesting machines worth $1.2 million in a plantation owned by Ekaterra, the producer of Lipton and Tazo teas. The clash resulted in two deaths and the arrest of Kericho’s governor. Ekaterra suspended operations for two weeks, leaving more than 16,000 employees without work.
One of the areas of dispute had been a provision in the industry’s collective bargaining agreement that workers, most of whom are women, would be kept on as machine operators. Labor leaders say multinationals have “flatly refused” to implement this.
“I do not condone the destruction of property. But these workers are now fighting back because the tea firms keep moving the goal posts,” Sang said.
Workers manufacture thread in a textile mill in Narayanganj, Bangladesh, on Sept. 2. Kazi Salahuddin Razu/NurPhoto via Getty Images
Ellingrud said 85 percent of jobs impacted by generative AI will be concentrated in four job categories: food services, customer service and sales, office administration, and manufacturing. The first three are dominated by women. Even in manufacturing, women like Nyunja are vulnerable compared with men, who have a higher chance of being retrained for roles involving automation.
The COVID-19 pandemic showed that women are more likely to lose jobs during periods of economic upheaval and are slower to return to the workforce, Ellingrud said. “Women’s unemployment is more sticky.”
Her research at McKinsey has found that 12 million people will need to switch jobs by 2030 and that women are 1.5 times more likely than men to have to change their occupation. She said this means governments and businesses need to urgently take targeted actions to re-skill and up-skill women.
Automation has also radically changed the makeup of Bangladesh’s garment industry, once hailed for transforming women’s employment prospects. Women once made up more than 80 percent of the garment workforce; today, they account for less than 60 percent. In 2019, the Bangladeshi government projected that half a million garment workers, mostly women, would lose their jobs to automation.
At Moni Garment and Training Center in Savar, a garment hub northwest of Bangladesh’s capital of Dhaka, Mizanur Rahman drills his trainees on how to operate machines used for weaving and knitting. A former garment worker himself, he said being sensitive to small areas of concern goes a long way toward making women feel more comfortable.
This includes hiring female instructors for female trainees and offering flexible hours so that women can show up before or after taking care of household chores. He said the confidence women gain from their training can lead to increased recognition at work.
“Many of my trainees perform well and are promoted to supervisor or line chief positions,” Rahman said.
Bangladeshi women work in a garment factory in Narsingdi, Bangladesh, on July 27, 2009. Munir Uz Zaman/AFP via Getty Images
Supervisory roles and roles operating automated technology are likely to be among the key jobs that survive in a post-automation landscape. Both tend to be dominated by men. In Bangladesh’s garment industry, women have long made up less than 5 percent of supervisors despite being a significant majority of the workforce.
There have been some signs of success. The Gender Equality and Returns project, run by the World Bank’s International Finance Corp. and the United Nations’ International Labour Organization, said 60 percent of its trainees have been promoted to supervisory positions, and the number of female supervisors in the industry has jumped to 12 percent since the program began in 2016.
Abdullah Hil Rakib, a director at the Bangladesh Garment Manufacturers and Exporters Association, the country’s largest trade association for the industry, said the key obstacle to women thriving in supervisory or machine operator roles is psychological.
“It’s a barrier in our mindset,” he said. He pointed out that automation means less physically strenuous work for both men and women, eliminating one barrier that would have previously made some jobs less accessible for women.
Kenyan workers pluck tea leaves using a machine at the Unilever Tea farm in Kericho, Kenya, on Oct. 10, 2004.Thomas Mukoya/Reuters
“Even when a man runs a heavy automatic cutter machine, he only pushes a switch on and off. He does not need to do more,” Rakib said.
Ellingrud said about 10 percent of jobs created every year tend to be new roles that didn’t exist before, but women take up these jobs at a lower rate than men. People won’t be losing their jobs to AI, she said, but to people who know how to use AI; women who are unable to adapt risk being left out of the new economy.
Adaptation feels like a distant prospect in Kericho, where Nyunja hawks vegetables on the street to make ends meet.
“I used to be able to take care of my family and pay my children’s school fees,” Nyunja said. “Now my future looks bleak. I can barely pay my rent, let alone send my child to school.”
Maher Sattar contributed reporting from New York.
