The Hidden Stakeholders of Artificial Intelligence. How to Build AI Projects That Don't Harm People (en Inglés) - Gloria J Miller

Artificial intelligence is reshaping how decisions are made-in hiring, lending, healthcare, law enforcement, and many other areas of life. Yet when automated systems cause harm, responsibility is often unclear.

This book examines the people most affected by AI systems and least likely to be represented in project decisions. It introduces the concept of passive stakeholders-workers, customers, and vulnerable communities who may experience the consequences of automated decisions but have no influence over the projects that create them.

Drawing on five case studies and original research, the author presents a practical framework for AI project teams to identify affected stakeholders, assess potential harms, and establish governance structures that clarify responsibility across the full lifecycle of an AI system. The book introduces tools including a stakeholder responsibility workshop, a harm taxonomy, and a lifecycle-weighted scoring model.