Ethical Issues in Data Science

This module begins with an introduction to the course including motivation for the topic, the course goals, what topics the course will cover, and what is expected of the students. It then reviews the three ethical frameworks that are most commonly applied to ethical discussions in data science and computing: Kantianism/deontology, virtue ethics, and utilitarianism. Case studies are used to illustrate the application and properties of these frameworks.

This module begins with some background about the Internet, which is the foundation for most of the topics that we study in this course. It then discusses the two most basic ethical issues in using the internet, privacy and security, in the context of data science. It goes through a number of real case studies and examples for each to illustrate the diversity of issues.

This module provides insight into the ethical issues in the data science profession and workplace (as opposed to technical topics in data science). It starts with discussion of two highly relevant codes of professional ethics, from professional societies in statistics and in computing. It then looks at a variety of recent workplace ethics issues in tech companies. A key part of this module is interviewing a data science professional about ethical issues they have encountered in their career.

Algorithmic bias may be the topic that people associate most with ethical issues in data science. This module begins by providing some general background on algorithmic bias and considering varying views on the pros and cons of algorithmic vs. human decision making. It then reviews an illustrative set of examples of algorithmic bias related to gender and race, which is a particularly important class of instances of algorithmic bias. The final part of the module discusses what is perhaps the single most prominent and discussed instance of algorithmic decision making and bias, facial recognition.