Peter J. Welling III
Professor Frazier
Data 150
9 September 2021
Joshua Bloomstock explains that data science could transform international development by designing data-enabled applications that work in the real world and pay attention to the people during the process. To support Bloomstocks thesis, there are promising studies that support the idea. Studies show the same algorithms that google, facebook and other companies use to match advertisements to people can also be used in another way, to connect resources to people living in poverty. This is because they use digital signatures in personal data from mobile networks and in imagery from satellites. Although it does seem promising, there are pitfalls. One of the pitfalls is that there are unanticipated effects. Solutions from big data more times than not, lift up people who are already empowered, instead of the people who are vulnerable. Bloomstock uses an example of China, citizens were unable to use trains and aeroplanes due to poor credit. Another example is a 2015 study that was conducted found only 51 percent of borrowers were aware of the interest rate they were being charged, so it’s anticipated that many people from this study were not content with the loan they were given and therefore connected back to China’s example were not able to use public amenities. A second pitfall that was expressed was the lack of validation. In the example of people receiving aid or eligible to receive a loan, people have taken advantage of the system. A non profit organization called GiveDirect uses satellite imagery to detect people who need aid by looking at the foundation of their roofs. People soon caught on and pretended to live in homes with poor roofs in order to be eligible to receive compensation. Another pitfall that is expressed is biased algorithms. People who live in developing economies are often at a disadvantage due to the fact that digital data is in favor of wealthier areas because it is in use more. An example is some digital-credit firms require users to be connected to online resources as a way to connect, limiting the use for less wealthier environments. The last pitfall is lack of regulation. Private companies are mostly in control of receiving data from artificial-intelligence applications and their intent is to maximize profits which comes at a cost for the lower communities.
In order to try limit the pitfalls, Bloomstock lists several steps to try limit the concerns. The first way is through validation, where data sets should build off each other and not replace old ones. Another way is customization. This is where algorithms which we use for a specific task are now used for an alternative purpose, making it more valuable and resourceful. The last way is by deepening collaboration. Bloomstock explains it as inter collaboration between data scientists, development experts, governments, civil society and private sector, in order to achieve massive growth throughout the industry. Considering the idea of good intent is not enough in data science when dealing with people’s problems connects to the idea about how people used satellites to their advantage when receiving aid. From the pictures it appeared that people with roof foundations needed aid but people caught on and cheated the system for their own gain. Addressing transparency and how it’s an issue to many of these problems is true. Which is why having an increase in data research coincided with human based issues could lead to more progress and better results.
In data science the ideas of “good intent”, “transparency” and the “balancing act” are issues when regarding the further development of humans with data science. If improved upon, there would be visual benefits on both sides, for humans and for the development of data science. There are ways to improve upon this idea to become more effective and have bigger change. One way is for large tech companies that have access to these algorithms to invest their time into these areas with the incentives of local governments. Now this does come with downsides. Inorder for it to be effective, US laws should be in place to limit the amount of fraud. This would let poorer economic communities benefit while also benefit the expansion and use case of the technology being implemented by the tech firms.