Responsible data science

Course goals

On successful completion of this course, students:

• O1. familiarize with the state-of-the-art research on the ethics of data science
• O2. define, describe and recall basic concepts and principles underlying responsible data science
• O3. identify stakeholders and ethical implications of data science in healthcare, design, crime, education, science, job markets, business, journalism, etc
• O4. understand ethical implications of data science algorithms on privacy, surveillance, discrimination, access to information, security, free will, human rights, social norms, etc
• O5. write an essay and a critical review in the field of responsible data science.
• O6. work in a team to create a prototype for solving an ethical issue caused by using data science algorithms.

Assessment

• Remindo Exams 25 % (with proctoring, if not on-campus)
• Presentation of an Academic Paper 10%
  The paper is self-chosen from a given list. 
• Group Project
    (a) Group Presentation: Journalist Overview 2%
    (b) Group Presentation % Preliminary Prototype: Final Project 8%
    (c) Group Essay & Prototype 30 %
  Topic is self-chosen, though a recommended list is provided.
• Personal Essay 25% (+ up to 10% bonus, for exceptional essays only).
  The topic is a critique, both conceptual and technical, of a given group essay. 

For a retake the grade of the original test has to be at least a 4.

Content

Responsible Data Science is examined through the lens of 4 introductory dimensions:

1. Data Dimension
2. Algorithm Dimension
3. Human Dimension
   (a) Psychology of Human Biases
   (b) Ethics / Moral Philosophy
4. Design Dimension
   (a) Data Visualization and Interaction Design
   (b) Explainable Artificial Intelligence (XAI) 

In this course, students follow lectures and workshops,  read literature, engage in class discussions, give presentations, critique, and write an original essay on a topic related to a (self-chosen) real-world ethical problem related to data science in a particular domain. The project also contains a practical solution to the problem illustrated in a low-fidelity prototype. 

Format
6 hours/week  in total of either lectures or workshops (distributed in 2 days).
The  format  (i.e. online, hybrid or in-campus) will be determined close to the registration period and is subject to change. In all cases, lectures are not recorded (due to the nature of the course), thus time-sensitive attendance is necessary.