Project Principal Investigator(s): Bidisha Chaudhari, Janaki Srinivasan, Sachit Rao
The application of Machine Learning (ML) techniques in domains such as, health, education, urban development, and governance, has increased in recent times, with the aim of making public-service delivery more efficient. This increase is primarily due to the availability of “big data” and the ability of ML techniques such as self-learning algorithms which can recognize patterns and predict outcomes from this data. Recent studies on the use of these techniques on the ground have raised concerns about a. fairness and biases encoded in the data used to train these algorithms and b. the opacity of how algorithms arrive at their results. These concerns are of particular importance in public services as they have been shown to impact diverse sections of the population differently, mostly adversely for vulnerable sections.