Ben Green’s new book gives data scientists a roadmap for social justice

Data scientists are often urged to consider the ethical implications of their work. From biased algorithms to tools that reinforce inequalities, the potential harms are clear. Even so, practical guidance for data scientists on what to do differently can be scarce. 

This October, University of Michigan School of Information assistant professor Ben Green published a new book with MIT Press. “Algorithmic Realism: Data Science Practices to Promote Social Justice” offers a step-by-step approach to developing data science projects that advance social justice. 

For Green, the book grew out of conversations with data scientists who wanted to work responsibly but felt unsure of how to put those intentions into practice.

“Algorithmic Realism is trying to bridge what is often a pretty stark divide between the worlds of data science and the more critical, humanistic worlds that critique data science and point out the limits and harms of algorithms,” Green says.

The questions are also rooted in his own career. Initially drawn to data science as a way to make a social impact, Green found that the challenges he encountered often extended beyond mathematical models and technical design. They involved politics, institutional policies, power and whether a system would be useful to the people using it.

“If data scientists are going to build tools for social good effectively, we can’t just apply the skill set and methods we have,” he says. “We need to pretty fundamentally rethink what our methodology looks like.”

Green began working on the book in summer 2022, building on an earlier paper co-authored with his wife, U-M law professor Salome Viljoen. The book develops those initial ideas into detailed guidance for the full life cycle of a data science project.

Written primarily for data scientists and those who work with or manage them, the book also explores how data science education can prepare students to put social responsibility into practice. That approach draws on Green’s teaching at UMSI.

For a broader audience, Green hopes the book offers a vision of AI development beyond a choice between embracing today’s approach and rejecting the technology altogether.

“We can pursue innovation with AI that is actually good for society,” he says. “But it’s going to be very different in terms of the structures and methods compared to what is dominating the field today.”

RELATED 

Purchase and read “Algorithmic Realism: Data Science Practices to Promote Social Justice” at MIT Press. 

Learn more about Ben Green’s research and teaching by visiting his UMSI faculty profile.