Promise: Joshua Blumenstock believes that phones can be the key to international development. He talks about machine-learning algorithms and their abilities to gather data from people that otherwise would be too costly and difficult to reach.
Pitfalls: For one, unanticipated effects. Bluemenstock highlights the risks of misappropriation, and the fact that citizens may not be at benefit in some cases. The second pitfall is the lack of validation: common types of data-collection have been around for a long time now and have had the ability to be improved and properly examined, whereas the new technological approaches have not had the chance to adequately be assessed prior to use. Blumenstock worries that algorithms may not remain accurate over time and therefore project data that is different from reality. Another concern is the humankind’s tendency to always look for shortcuts, and therefore the worry of people trying to cheat the system becomes more significant. Another pitfall that Blumenstock describes is biased algorithms. He explains that since data would be collected through phones, and possession of a phone can somewhat directly be tied to a person’s income, the poor segments of the population would be understudied and misrepresented in the data. The fourth pitfall that Blumenstock describes is the lack of regulation. He talks about private companies being in control of this data collection process, which raises the question of privacy and regulation.
Ways forwards Blumenstock uses as foundation for thesis:
Response to Anna Raymond’s “good intent” thought: I agree that having good intent is not enough since that can become an easy excuse. People working on the new data collection technology must have the geographical location and real people in mind, as opposed to treating it like a computer simulation.
Response to Nira Nair’s “transparency” thought: Personally, I’m a big believer in honesty and transparency no matter what the issue at hand is, so I couldn’t agree more that these data collections should be more open. Since it no longer would be a government task, the collection process and data in general should not be treated as classified information.
Response to Kayla Seggelke’s “balancing act” thought: I’m not sure I understand what is meant by the word “application”, but from what I think of the balancing act between being respecting people’s privacy and yet being transparent about the results, I believe that what Joshua Blumenstock was saying about deep collaboration is truly key. If there are people from various institutions and backgrounds working together on this project, I believe that it provides the best chances for success.