“At Microsoft, we say we want to empower the entire planet, right? And more than half the world’s population uses languages other than English.”
Bali says that AI has greatly sped up the process of language preservation and its use in large language models (LLMs). This is useful in creating online and AI tools, but also for preserving rare or dying languages.
“Now we can create these copilot kinds of things really quickly,” she says. “Previously when we were talking about language preservation, we were talking about efforts that took place over decades, literally. … All of that can now be shortened to months.”
Karya, which says it is on pace to engage with more than 100,000 workers by the end of 2024, seeks participants who need work and education the most – often women in rural areas. In addition to a premium wage, it offers training and other kinds of support when the work is done.
‘Technology can really, really help amplify people’s desires’
Chopra grew up in a “basti” – an informal settlement – in Delhi and says that the inequities he witnessed growing up had a profound effect on his sense of purpose while studying computer science with a focus on AI at Stanford University in California.
“When I moved back to India, the first thing I realized was, everywhere I went, people had the intent or the will to get out of poverty, everyone works really hard, everyone has aspiration,” he says. “And they have the capacity to learn new skills. And if those two things exist, technology can really, really help amplify people’s desires, to make something of themselves.”
The work Bokale did over the course of 11 days was part of a pilot project to test whether the work of inputting data could be combined with the learning of useful information. While earning what for them was a substantial amount of money, they’d also be learning about the financial tools they need to make the best use of it.
The material was presented as a serialized story about two sisters, and it was this story that the workers read aloud into their smartphones to capture the sounds and rhythms of spoken Marathi. “We really enjoyed the story,” Bokale says, “And in that story, there were common people who are working hard every day. The money they earned would be easily spent, there were no savings. In short, the question was how to save.”
Safiya Husain, Karya’s chief impact officer, said that the story format proved a success, and that many of the participants read the story out loud to their families and friends.
“They would say, ‘I’m going to do this work and read the story to you,” Husain says. “And they would actually get excited and wonder, ‘Oh, what is happening next? Will she get her loan? Or will she have enough money to pay for the wedding?’”
She says that by combining work with education, Karya was trying to treat its workers with respect and create outcomes beyond income that are meaningful. “We were paying people for their time, and we were saying what they were doing was valuable,” she says. “It wasn’t just, here’s a lesson to learn in your spare time.”
Husain says she hopes that eventually many of the Karya workers will join the organization in different roles, working as organizers and local administrators. In the big picture, she says, the aim is to put technology to work for everyone.
“When we’re collecting data in these languages like Marathi, we’re trying to make sure that these communities and these populations, which have millions and millions of speakers, are not being left behind in the technology revolution,” she says.
Engaging whole communities in the project
Kalika Bali, the Microsoft researcher, says one of the keys to the success of Karya is that it strives to engage whole communities in the project. Most of Karya’s workers are women, and she says they have more “circles of trust” to cross than men.
“The men only need to ask two things: will this work for me, and will I get paid?” she says. “Women have to ask; will my family accept it? Will this bring a bad name to my family and myself by doing this? Is this going to harm me in some way? Only then does it come to the platform and the money.”
“The advantage with Karya is that it has created a lot of trust on the ground. They are really engaged with the communities they’re in,” she says.
In her neighborhood in Pune, Bokale is a well-known figure who is universally known as Baby Tai, tai meaning “elder sister.” She runs an informal financial network with several dozen other women who pool savings monthly and take turns taking a larger amount to use for things like starting a small business or paying school fees. Women often show up on her tree-shaded patio to talk business or just hang out. Her chili and spice grinding equipment is in a small tin shed on one side of the small yard.
Surekha Sanjay Gaikwad, 51, is one of her neighbors and friends. She runs a small grocery store about a half hour from her home. She also reads Marathi into her phone for Karya. Sitting with Bokale on her front steps, she burst into a wide grin when asked what she liked about the experience.
“I couldn’t believe I could do it at home,” she says. “I don’t have to get on a bus again or go anywhere else at the end of the day.”
The education component of the work was a plus, Gaikwad said. She learned how to create a fixed deposit at the bank, and she did just that as a way to save more effectively for her son’s college studies.
Over the course of a recent morning, several other women who had worked for Karya stopped at Bokale’s home to chat. Meena Jadhav, 55, had used the money to buy material and sewing tools for her tailoring business – she made shirts to sell. Thanks to what she learned, she said, she can now use a savings account and knows how to use an ATM. She didn’t know you could withdraw and deposit money without going to the bank.
Another woman used the lessons she learned and the cash she earned to start a savings account for her daughter’s education.
They all said they enjoyed the work and found the information about financial planning and online tools useful. An added benefit for the women, Bokale says, was learning that their smartphones could open doors to other kinds of opportunities.
She says many of the other women in the pilot project didn’t know how to use a smartphone at all beforehand. “Their husbands and in-laws, they’re saying ‘Oh, you’ve learned so many new things, and that’s so great.’”
Top image: Baby Rajaram Bokale with some of the women in her informal investment group. By recording and writing Marathi in their smartphones they helped create datasets to be used to create AI language models.
