Bluemenstock response Fatima Pate 01/28/2020

Blumenstock’s main argument is that big data should not disregard human development in its advancement. He begins by explaining the promises of data science. The recent increase in mobile technology has allowed for data scientists and major corporations such as Google and Facebook to track individuals “digital signatures” and “digital footprints” in order to provide the most efficient mode of help. For example, an individual in Africa’s digital footprint may indicate whether they are wealthy or poor and their living condition. Based on this, it will be easier to provide help during times of crisis or an epidemic.

Although these tools do have the ability to contribute positively to human development, they are imperfect. The four problems outlined by Blumenstock regarding the ways in which data science attempts to impact human development includes unanticipated effects, lack of validation, biased algorithms and lack of regulation. The first major issue is that “Solutions enabled by big data often bolster those who are already empowered.” Because the data tends to come from a select already powerful few. Secondly, small sets of data tend to be generalized and integrated into places where it ends up not having the same positive affect as it would somewhere else. Biased algorithms like the data collected from Google Maps and Waze discounts those who do not have phones and consequently benefits the wealthy who do own phones with the ability to carry these applications.

To further the progress of data science in human development, Bluemenstock believes it is crucial to first, validate data—“New sources of data should complement, not replace, old ones” meaning all data, whether old or new is essential to human progress. Secondly, customization is key to providing “fair, accountable and transparent” aid to people around the world. A model that works in the United States may not work in Nigeria and so it is important that scientists first understand the needs of a country, and base their method of help on the needs of the country rather than expecting that all countries respond similarly to the same product. Thirdly, there has to be more collaboration amongst data scientists, the government, development experts, and the citizens. Data scientists and algorithms can only do so much to tackle human crises but the work is essentially inefficient and useless if it is not integrated and analyzed by those directly affected by it and by those who specialize in understanding human development. I believe it is true that “good intent” can only take you so far as a data scientist involved in the well-being of humans around the world. “Good intent” is the first step but it must be followed by active engagement in the world and active efforts to better understand humans and their needs otherwise no legitimate progress will be made.

Regarding transparency, I completely agree with Nira. I believe it is critically important that both ends are willing to have an open line of honest communication and engagement with one another or consequently both sides will experience a plateau of progress.

Finding the balance between “promoting applications yet demoralizing hinderances” as Kayla put it, is hard. The difficulty in this comes from the fact that human development is a constant pressing matter which requires many forms of input, data science for example. Yet data and applications are required to deal with human development so it becomes difficult to balance out how much attention is given to both sectors.