Ethical Data Science

Abstract

Predictive data science is often viewed as a public good, yet too often human needs are pitted against the financial interests of large digital infrastructures. Can data science truly serve the public interest? The question of our time is whether the promise of predictive science will exceed the inherent conflicts of interests presented by the organizations that fund it. As one of the first books on public interest technology, Ethical Data Science: Prediction in the Public Interest acknowledges the institutional incentives that drive computational prediction. It argues that data science prediction embeds administrative preferences that often ignore the disenfranchised. The book introduces the prediction supply chain to highlight moral questions alongside the interlocking legal and commercial interests dominating data science. Situating data-driven analysis within multiple layers of effort systematically exposes risky dependencies while also pinpointing opportunities for principled action, research ethics, and policy interventions. The prediction supply chain organizes critical thinking throughout the typical data science workflow, from gathering data through learning from insights. Data scientists can explore the full human potential of prediction by thinking like chefs, painters, historians, librarians, philosophers, and architects. Each stage draws on the arts and humanities as methods to empower broader conversations across multiple forms of expertise. Practitioners, academics, students, and legislators will learn how to identify social dynamics in data trends, reflect on ethical questions, and deliberate over solutions. The book establishes the limits of predictive technology controlled by the few and urges for more inclusive data science.