Although these two approaches produced similar estimates in this case, the approach of Blumenstock and colleagues was about 10 times faster and 50 times cheaper than the traditional Demographic and Health Surveys.
This shows how the same data can help public policy but also enable tracking people. Even if the researchers are careful, someone else could use similar methods for harm. It makes me think ethics should include thinking about misuse, not just the study’s intention.
Combining these two sources of data, they used the survey data to train a machine learning model to predict a person’s wealth based on their call records.
The survey gives “real” wealth info, and the call records give scale. But what phone behavior is really measuring?
researchers can now observe the behavior of millions of people, and as I’ll describe later, researchers can also enroll millions of people in massive experiments.
This is a helpful reminder that big data isn’t enough by itself. Traditional methods (surveys, interviews, careful sampling) still matter for meaning and validity. The best work is probably a mix, like the Rwanda example.
Further, it can easily randomize groups of customers to receive different shopping experiences. This ability to randomize on top of tracking means that online stores can constantly run randomized controlled experiments. In fact, if you’ve ever bought anything from an online store, your behavior has been tracked and you’ve almost certainly been a participant in an experiment, whether you knew it or not.
Even if A/B tests are “normal,” it raises consent and transparency issues: does “implied consent” exist when users don’t know they’re being experimented on? It also makes me think about the boundary between product testing and research. At what point does an A/B test become human subjects research?
The fact that people might respond emotionally, however, does not mean that you should dismiss them as uninformed, irrational, or stupid.
This quote reminds me how often experts dismiss criticism by calling it “emotional” or “uninformed.” It’s important to always remember / acknowledge that emotional reactions can reveal ethical blind spots.
That is, a database created for one purpose—say targeting ads—might one day be used for a very different purpose
This idea of unanticipated secondary use is unsettling. How should researchers today anticipate potential misuse decades later, especially in unstable political climates?
rather than imagining yourself in the watchtower, imagine yourself in one of the cells.
This metaphor flips the researcher's perspective in a haunting way. It helps me understand how even well-meaning data collection can feel invasive or oppressive to those being observed.
This principles-based approach helps researchers make reasonable decisions for cases where rules have not yet been written, and it helps researchers communicate their reasoning to each other and the public.
This shows why rules aren’t enough in fast-changing tech environments. The idea that ethical reasoning should be communicated. It adds a layer of transparency to the ethics process.