Karya

Karya operates as a non-profit, ethical data company with a mission to provide digital work to rural Indians, creating a pathway out of poverty. Established in Bengaluru in 2021, the company was founded by Manu Chopra, Vivek Seshadri, and Safiya Husain, who joined a year later. The founding concept originated from research conducted by Chopra and Seshadri at Microsoft Research, exploring how to offer digital work to supplement the physical labor provided through government employment schemes, leveraging India's increasing smartphone penetration.

Co-founder and CEO Manu Chopra, a Stanford University computer science graduate, was motivated by his firsthand observation of the exploitative low wages in data work, inspiring him to create a more equitable model. Co-founder and CTO Vivek Seshadri is a Principal Researcher at Microsoft Research India with a Ph.D. in Computer Science from Carnegie Mellon University, focusing on leveraging technology for global development. Safiya Husain, the Chief Impact Officer, brings a background in humanitarian work, research, and evaluation, having grown up in post-conflict regions and worked with organizations like the UN and STiR Education.

The company's business model is centered on providing data services to major technology clients, such as Microsoft and Google. Karya takes on large-scale AI data projects—like generating thousands of hours of audio data in specific Indian languages—and breaks them down into microtasks. These tasks are distributed through Karya's mobile-first digital platform to workers in rural communities. The platform is designed for accessibility, offering offline capabilities and vocal assistance for users with limited literacy. The business generates revenue by selling these high-quality, ethically sourced datasets to its clients. A key differentiator is its compensation structure; Karya pays its workers significantly above the Indian minimum wage, often around 20 times higher, and offers royalty payments when datasets are resold, ensuring that a fair share of the value generated by AI is returned to the data workers.

Karya's service involves capturing, labeling, and annotating data with high accuracy, specializing in creating datasets for low-resourced Indian languages that have been previously marginalized in the AI boom. This focus not only provides crucial data for building more inclusive AI systems but also creates dignified, supplementary income for its workers. The company partners with on-the-ground non-profits to recruit workers from marginalized communities, ensuring access is distributed equitably based on income and diversity requirements. The work is designed to be simple and accessible, such as recording sentences in a native language on a smartphone. This model has enabled Karya to distribute millions of tasks to tens of thousands of people across numerous Indian states, gaining recognition on the cover of TIME Magazine and endorsement from Microsoft's CEO, Satya Nadella.

Keywords: ethical AI, data annotation, data labeling, rural employment, social impact, AI training data, Indic languages, microtask platform, poverty alleviation, digital livelihoods, fair wages, impact sourcing, data cooperative, AI for good, language data, human-in-the-loop, Microsoft Research, inclusive AI, community empowerment, mobile work