Advanced Topics in Computer Science: Responsible AI in Societal Deployment

What responsibilities, and for whom? This course addresses the theoretical and sociotechnical foundations of responsible AI. We survey current approaches, as well as the philosophical, ethical, and legal principles shaping AI in societal deployments. Methodological aspects include: data choice, problem formulation, prediction modeling limits, interventions/causal inference, measuring/mitigating algorithmic biases, algorithmic fairness limits, and impact evaluation via experimental design. We emphasize understanding broader impacts, relating to social justice, via case studies from education, healthcare, financial and social services.