Courses
STOR 323 -- Machine Learning: Ethics and Society

Ethics of Machine Learning in Society

Course Overview

This course examines the ethical, social, and technical challenges arising from modern machine learning (ML) and artificial intelligence (AI) systems. It integrates

  • technical foundations including supervised learning, online learning, mult-agent modeling and game theory
  • philosophical frameworks such consequentialism, deontology, virtue ethics

to analyze the impact of real-world ML applications such as recommendation systems and generative AI.

The course emphasizes the tension between algorithmic decisions and societal impact, training students to critically evaluate and design responsible ML systems.

Logistics

  • Instructor: Fan Yao
  • TA: Xianwen He (xhe1+at+unc.edu, Hanes B33)
  • Lectures: Mon/Wed/Fri 12:20 – 1:10 pm, Hanes 130
  • Office Hours: (Instructor) Mon 2-3pm, Hanes 338, (TA) Wed 3-4pm, Hanes B33

Learning Goals

  • Analyze ethical dilemmas in ML/AI systems
  • Understand technical tools used to address these issues
  • Evaluate real-world deployments and their societal consequences
  • Communicate arguments through writing, presentations, and discussion

Course Structure

The course is organized around three main themes:

1. Moral Pitfalls of ML

  • Intro to Moral Philosophy
  • Foundations of ML (classification, regression, overfitting)
  • Bias, fairness, privacy issues in ML
  • Online learning and feedback loops in ML

2. Social and Economic Implications

  • Human factors in ML systems
  • Game theory and strategic behavior
  • Social choice, equilibrium, and mechanism design
  • Applications to platforms and recommender systems

3. Generative AI: Challenges and Opportunities

  • Generative AI (GenAI) and Large language models (LLMs)
  • Aligning LLMs with human values
  • Ethical and societal implications of GenAI

Full syllabus and course materials available on Canvas. For visitors outside UNC, course slides are available upon request.

Assignments and Evaluation

  • Homework (20%): Technical analysis + ethical arguments
  • Quizzes (20%): Short conceptual in-class assessments
  • Participation (10%): Discussions and peer engagement in class
  • Midterm Proposal (20%): Identify and analyze a potential ethical issue brought by modern ML/AI applications
  • Final Project (30%): Design and analyze a technical or policy intervention

Course Design

This is an interactive, discussion-driven course rather than a traditional lecture-based course. Students are expected to engage in:

  • Structured debates on real-world ML systems
  • Case studies of ethical dilemmas
  • Team-based projects addressing contemporary issues

The course encourages collaboration, critical thinking, and practical intervention design, preparing students to navigate the ethical challenges of deploying ML in society.

Tools and Format

  • Programming: Python
  • Writing: LaTeX (Overleaf recommended)
  • Materials: Open-source readings and lecture notes
docs