DSC 291: Data Science, Ethics, and Society (Winter 2022, UCSD)

Published:

Note: This is a draft syllabus, some details may be subject to change.

Details

  • DSC 291: Data Science, Ethics, and Society
  • Profs: David Danks (HDSI/Philosophy) and Stuart Geiger (Communication/HDSI)
  • Time: Tu/Th 6:30-7:50pm
  • Place: WLH 2207

Overview

Data and AI are rapidly influencing all parts of society, with corresponding rise in the need to understand the ethical, personal, and societal impacts. This course will cover issues of privacy and consent; fairness and bias; power and justice; explainability, transparency, and interpretability; trust, accountability, and contestability; and how these topics can be put into practice with social and technical mechanisms. Students will learn to identify ethical & societal opportunities and risks in data science and AI in both their own and others’ projects. A final project/analysis is required in one of several possible formats (term paper, website, presentation, interactive code or visualization).

Learning outcomes:

  • Understand core ethical and social scientific concepts that apply to data science, including privacy, fairness/bias, power, explainability, and justice
  • Use these concepts to identify ethical & societal opportunities and risks within a data science project
  • Derive specific (potential) ethical & societal impacts of standard data science practices, methods, and products
  • Develop both ethical and social scientific analyses of the permissibility and/or legitimacy of a specific data science project

Schedule

  1. Foundations
    1. What is ethics? What is social science?
    2. Introduction to key methods used to answer ethical & social scientific questions
  2. Power
    1. Core concepts & considerations
    2. Historical & social sources of power imbalances
    3. Data science as vehicle for maintenance of hierarchies of power
  3. Fairness/bias
    1. Core concepts & measures
    2. Ethical & societal considerations
    3. Grounding technical choices in substantive theories
  4. Justice
    1. Core concepts & considerations
    2. Historical & social manifestations of injustice
    3. Data science as vehicle to reduce injustice
  5. Privacy
    1. Core concepts
    2. Ethical & societal considerations in practice
    3. Key legal distinctions and practical (psychological) guidance
  6. Explainability/Transparency/Interpretability
    1. Core concepts & considerations
    2. Determining the features that are required in a particular case
  7. Trust/Accountability/Contestability
    1. Core concepts & considerations
    2. Methods to build (and restore) trust, accountability, and contestability in data products
  8. Putting the concepts into practice / user/community-centered design
    1. How do we actually find the right concepts & questions in a real case? (e.g., how can we decide if privacy considerations (or others) are actually relevant in this particular situation?)
    2. And how do we answer these questions? User/community-centered approaches for including stakeholders in design