Ramapo invites speaker to examine the ethics of data science

Photo courtesy of Lala Ros, Flickr

Ramapo College hosted Dr. Andrea J. Pitts from the University of North Carolina at Charlotte to present “Data Ethics to Data Justice: The Challenge of Building Better Worlds through Data Science” on Feb. 22. Pitts claimed data ethics is expected to turn the current way the world looks at data onto its head with the idea that “assessing people as knowers can become a moral issue.” Over time, the process of determining credibility will start to include ethical factors such as how the data was collected.

Pitts earned a degree in philosophy from Vanderbilt University, is an assistant professor in the subject and serves as an affiliate faculty member at the UNC Charlotte School of Data Science, where she teaches students how to use machine learning, data analysis and artificial intelligence.

She has a wide background in biomedical ethics, social epistemology and critical philosophy of race and gender. Now, she is working to implement data ethics, an emerging interdisciplinary field stemming from media ethics, into her university’s data science program.

Danah Boyd and Kate Crawford are often credited as the founders of the emerging field. Pitts considered their arguments, “big data changes the definition of knowledge” and “claims to objectivity and accuracy are misleading,” to be the roots of a paradigm shift within the field of data science as a whole.

As big data changes what information is seen as more important or reliable, more ethical considerations are needed to ensure this information is being collected and used fairly. If big data can contribute to changing labor practices, and AI can influence who gets targeted by police investigations, organizations like the Algorithmic Justice League are important to ensuring marginalized groups are not being targeted.

Constant examinations of the ethical, social, cultural and political aspects of data science have suggested “the numbers don’t speak for themselves” and bigger data is not always better data, depending on how it is collected. For example, if the subject of a study is all mobile phone users, scraping Twitter and Facebook for data will exclude all mobile phone users who do not use those platforms.

All of this has inspired Pitts to arm new generations with the skills to ask and answer ethical questions about how data is collected and used. She and her colleagues, Professor Stephanie Moller and Dr. Angela Berardinelli, developed four courses corresponding with data ethics student learning objectives: Data and Society, Modeling and Society, Predictive Analytics and Their Implications, and Data Science for Social Good.

The courses promote “non-individualizing” approaches to data collection and expand upon the time-honored phrase “Just because something is possible doesn’t mean it should be done… ought to be used… (or) made public.”

Pitts admitted collaborating across several colleges within UNCC to implement the courses within the existing data science program was no picnic, with challenges ranging from limited time to beginner students lacking the skills for analyzing larger data sets. However, she considers the effort to be worth the cause. 

Data science is “shaking the social landscape,” and Pitts wants her students to have a finger on the pulse of the epicenter.

 

dbongiov@ramapo.edu