Models and data-fueled applications are everywhere, and thus that is exactly where data science skills and opportunities are needed — everywhere. This is the deceptively simple but powerful idea behind Responsible Data Science@Pitt: Data Science is about people — and thus we cannot stop at the screen and silicon.
Too often data science is shielded and siloed by distance and abstraction from any responsibility for the harms caused by gaps in the available data or convenient but distorting modeling assumptions. Too many data science programs acquaint students with the context of data from the Titanic and Sepal pedal length (two famous introduction to data science example data sets) but leave them blind to ways data and modeling can be used to lead and also mislead in the relevant fields where they will work and/or research. This gap leads to very real on-the-ground costs and increasing disparities in digital access and agency. At RDS@Pitt, we are broadening the focus to responsible uses and applications of the ever-evolving tools of data science.
Our approach reflects the growing acknowledgment of two facts across thought leadership in data science.
First, that the positive influence of data and modeling emerges from their responsible use by people and teams in every industry imaginable. Everything from archiving to astronomy, logistics to leisure, manufacturing to molecular biology, and from cancer research to coaching has been changed by the availability and processing of data.
Second, there is not currently — nor will there ever be — a one-size-fits-all curriculum for building and deploying responsible data science tools within specific contexts. Every application in each industry requires thinking through distinct tradeoffs and values. There are common lessons across industries and research domains to share, but they require reasoning and communicating beyond the data and model.
Because of the importance, complexity, and scope of these challenges, we cannot be timid in our training and engagement in responsible uses of data science. We need to engage not only the STEM of data analysis but also the more difficult to see human roots — where the data comes from and who decides — and the human fruits — who will be influenced and in what way.
With this in mind, RDS@Pitt is organized on a whole-University approach to pioneering new training and research programs. Our goal is to network together and empower people in every field and from every background — be they traditional undergraduate or graduate students or workers who have not been inside a classroom in two decades — to steer data and models in ways that maximize their positive uses and minimize harms.
Make no mistake, if we do not prepare many more citizens and workers now and in the future to understand and direct data analytics and mould modeling choices, we will continue to see avoidable community harms from data science while the benefits accumulate to the few.
Please do not hesitate to reach out to us if you have ideas for a Community of Practice, Working Group, or event related to your broad mission at RDS@Pitt. We are eager to partner with you and amplify the incredible energy in responsible data science that already is bubbling up across Pitt.
— Mike Colaresi, Associate Vice Provost for Data Science