AI as a force for good

April 29, 2019
A Mumbai-based nonprofit research institute is tasked with developing AI for social good.

This is an extract from “The ecosystem“, the first report of the four-part series, “Asia’s AI agenda”, by MIT Technology Review Insights.

P. Anandan, is an AI scientist and the CEO of Wadhwani Institute for Artificial Intelligence, a Mumbai-based nonprofit research institute tasked with developing AI for social good. He sees Wadhwani’s mandate—and all AI actors in India—as being to identify the roots of the country’s human development challenges and focus AI on specific solutions.

“The greatest tangible impact AI can have on [India’s] economy and industrial landscape is not about process automation, it is about core productivity. The Indian economy is still very driven by human labor, and with 1.2 billion people, that’s not likely to change soon. Making those people optimally productive is our biggest challenge.” A key area of economic productivity is India’s huge agricultural economy, which still employs half of all workers in India. Anandan’s team has been working on AI tools for cotton farmers in Maharashtra, which are helping them tackle such productivity challenges as pest management.

“The greatest tangible impact AI can have on India’s economy and industrial landscape is not about process automation, it is about core productivity.”

P. Anandan, AI scientist and the CEO, Wadhwani Institute for Artificial Intelligence

Healthcare is fertile ground for Wadhwani. Anandan describes how 900 million people are served through a national network of 25,000 primary health care facilities, with a further 100,000 sub-centers in rural areas that do not have permanent staff. Nurses or doctors visit once cases have been referred upward. Wadhwani worked with a sister foundation, the Wadhwani Initiative for Sustainable Health (WISH), that had experience in rebuilding, redesigning processes, and running 100 subcenters in Rajasthan for two years before handing them back over to the government. In redesigning the processes, the team identified a dozen use cases where AI would help. One was a triaging app for healthcare caseworkers around the identification of risks associated with pregnancies and prenatal care, for prioritizing the locations and sequencing of their field visits. The apps will soon be available for field trials by health and wellness centers on a larger scale.

Anandan sees infrastructure as India’s second root challenge. “Public utilities and transportation infrastructure—roads, railways, waterways—can be more efficiently managed and better planned.” The third one “which has always been very important in India, is actually e-commerce itself. We already see middle class and even lower-middle class households in India relying on online mechanisms for doing a number of their everyday purchases.” Making digital commerce more efficient could unlock considerable value in the economy.

While there are many daunting challenges, including India’s lack of deep pools of structured data, he is positive on the outlook. “None of them are insurmountable. The kind of decisions, about data collection, will take some time for us to figure out—actually, just gauging the infrastructure on that scale in order to collect and share all the data, and deal with privacy and security issues does create many, many obstacles. But if we start with specific problems and design solutions on a small scale, on a scale of our healthcare initiatives, we can use them as pilots with which we can succeed at a district level, and eventually follow through that scaling process to a national level.”