As part of UC Berkeley’s virtual conversation series, campus faculty discussed ways to understand and seek equity during the coronavirus pandemic.
Associate Provost of the Division of Computing, Data Science and Society Jennifer Chayes hosted the livestream with campus researchers, including acting associate professor of health policy and management Ziad Obermeyer, School of Information assistant professor Niloufar Salehi and anthropology professor Sarah Vaughn. During the livestream, each researcher presented an overview of current studies in relation to the coronavirus pandemic.
According to Obermeyer, bias involved in algorithms and data collection has the potential to be “dangerous” both generally and when looking to understand the pandemic’s severity. Algorithms require specific variables and datasets that bring to question how the “messy” concept of health and deterioration was translated into a specific variable.
“Be careful what you ask an algorithm to do — it’s going to do it,” Obermeyer said during the livestream. “If you ask it to predict health care costs that’s what it’ll do. It won’t pause to ask if cost is biased because of historical disparities, it won’t point out that Blacks are underrepresented in the highest risk groups — that’s on you.”
He added that when making COVID-19 projections, data collection can also be problematic. Access to health care is needed to get tested. Without it, Obermeyer said positive cases won’t be included in data, adding that correcting these choices can make a “huge difference.”
Salehi’s research is focused on marginalized groups’ use of social systems and the ways in which technological advances can better suit human needs. According to Salehi, working with these communities requires careful consideration of research methods and data sources.
“We did a lot of work to build relationships and trust with community leaders before we even started to gather data,” Salehi said. “This is to ensure that the research serves their interest and that they are partners in the work, not just research subjects.”
Salehi added that the idea that the Black and Latinx community’s high mortality is due to “their own risky behavior” is “dangerous.” According to Salehi, these remarks can be attributed to a long history of systemic accreditation to individual shortcomings. Salehi said researchers must be aware of the role research has played in creating this narrative.
Vaughn discussed her climate adaptation work, where she collaborates with engineers to understand these issues when living with uncertainty surrounding infrastructure. She added that those living next to critical infrastructure recognize that these systems need to be improved and repaired.
“Climate adaptation is such an interesting case by which to think about a comparative way to get into questions of COVID-19 … precisely because of this issue around living with uncertainty,” Vaughn said during the livestream.