80-832 Seminar on Ethics and AI: AI Value Alignment
12 units, Carnegie Mellon University
This graduate course explores the complex landscape of AI value alignment from complementary ethical and safety perspectives. The course examines how different philosophical and ethical theories inform our understanding of what constitutes ""alignment"" while engaging with technical approaches that attempt to operationalize these concepts in AI systems. Throughout the semester, we will investigate tensions between idealized ethical theories and practical implementation constraints, analyze how safety concerns shape alignment research priorities, and consider the sociopolitical dimensions of defining whose values should inform AI alignment. Students will engage with diverse literature spanning machine learning, moral philosophy, philosophy of science, and AI governance, cultivating a multidisciplinary understanding of alignment challenges that transcends purely technical or purely philosophical framings. The course aims to foster critical constructive thinking about the fundamental assumptions underlying different approaches to alignment while considering both near-term applications and longer-term implications as AI capabilities continue to advance. The purpose of this course is to provide a solid introduction to the risks of artificial intelligence, the challenges of controlling AI systems, and the alignment problem as a central concern for developing and governing AI responsibly. This course can be considered as a requirement of ethics and AI PhD certificate.
Prerequisites: none listed
Offered: Fall 2025
Courseweave needs JavaScript to show sections, workloads and planning tools.