This is a user generated content for MyStory, a YourStory initiative to enable its community to contribute and have their voices heard. The views and writings here reflect that of the author and not of YourStory.

Can machine learning help tackle development problems?

The scope of ML is very expansive in a wide variety of development indices.

Can machine learning help tackle development problems?

Wednesday August 01, 2018,

4 min Read

The adaption of technologies into the field of human welfare is not new. For decades, space technologies have been used to study patterns in crops. This continues to be a significant activity in many developing countries, where satellites are being used for development in rural areas. Is it possible to extend Machine Learning (ML) to help tackle development problems? Can it go farther than existing technologies in tackling development issues?


The World Bank seems to think so. At this stage, it may not be possible to say if Machine Learning will make the world a better, more developed place to be in at some specific point of time, but like all technologies, ML too is going to take its time to get implemented and adapted in different areas of development.

The World Bank, in a Knowledge for Change Program (KCP) project study it carried out in early 2018, tried to analyze how ML can be used to evaluate and predict development outcomes more positively and accurately. The study found that ML, when fused with AI, can make a palpable difference in these areas of identifying and predicting parameters of development:

Measuring changes: The most important aspect of development that ML [Machine Learning] was found to foster and hasten understanding of when it comes to the effectiveness of development programs, is in its ability to help measure changes. Several interventional programs keep getting implemented by the governments of developing countries (in this case, the countries chosen for this study are those in sub-Saharan Africa.) The study found that ML is more effective than traditional, barefoot physical methods in zooming in on the exact levels of development that the interventions have sought to achieve. For example, it throws up keener understanding of urban indices such as building heights and specifications, which could help check tax defaults or encroachments.

Sharper focus on treatment: The scope of ML is very expansive in a wide variety of development indices. These go beyond understanding or evaluating aspects of rural development. For instance, ML can facilitate far more effective tax collection on goods passing through very heterogeneous geographies and markets such as India. It could also help policy makers locate exactly which areas of a region are more affected by deforestation, and so on.

ML and AI as part of the solution: While using ML and Artificial Intelligence as tools that bring about greater and more insightful understanding of many problems relating to development; the trick, if you like, consists of using these technologies as part of the solution. ML promises to be an intuitive and specific solution to problems relating to development, according to this study. Used for areas as varied as immigration checks to helping farmers understand the best price for their yield; ML has the potential to transform lives by making development programs far more effective than they are today.

Promises to herald huge changes: This means a sea change in the lives of people in developing countries. There is usually no dearth of development programs in most of these countries, but most of these are on paper, being severely hampered by an inept bureaucracy and lack of will in implementation or accountability. The ultimate solution, which ML promises, it so becomes to the solution itself, by carrying out the objectives all by itself, rather than aiding in decision-making. We should ideally be in a situation where ML can be used to actually implement the solution after having identified it. But is that possible in the near future?

Given the level to which ML has developed at this point of time, one can say that a solution of this extent and reach is somewhere between being science fiction and becoming a future certainty. Further, even after ML technology could grow to a stage where it could actually make all these happen; implementing the solution requires a colossal change in thinking on the part of the governments. It may take quite some time to see changes at this high a level. With ML showing the way and promising to be the most powerful facilitator for change, or for it to actually deliver on its promise, could be some time away. 

Montage of TechSparks Mumbai Sponsors