Claudia Scholz, Director of Research Development at the School of Data Science, was a featured speaker last month at Howard University’s virtual series for Karsh STEM Scholars. She spoke about how data is acquired and the ethics behind these processes, as well as how data science impacts race and bias.
Scholz discussed how natural language processing costs, digital footprints, data dissemination, and machine decision-making power actually impacts the environment and at what cost.
“Because of the hardware, the server farms, the electricity, and the cooling infrastructure, the natural language processing model creates as many carbon emissions and climate impacts as driving your car for an entire year,” Scholz explained. “We have to ask ourselves if the analysis we want to run is worth this much pollution.”
Scholz also discussed the human assumptions involved in artificial intelligence and machine learning. “When we look at a data set, we run the risk of replicating and deepening human biases, including racism in electronic decision making.”
The Karsh STEM Scholars Program supports high-achieving students from STEM backgrounds interested in pursuing a PH.D. with an emphasis on research or teaching.
As the Karsh Scholars look towards careers in research or teaching in STEM fields, Scholz encouraged the students to always think about bias and ethics implications in all that they do, and what machines do.
“We might not think about ethics and computers all that much, but it is really central to understanding how data is being used and how it is going to shape your own lives and your careers.”
Claudia Scholz, Director for Research Development at the University of Virginia’s School of Data Science, earned her doctorate in sociology from Johns Hopkins University. She supports School of Data Science faculty and staff research and is a founding co-chair of its Diversity, Equity, and Inclusion Council. Prior to joining the School of Data Science, Scholz served as Associate Provost for Research at Spelman College.