Postgraduate Programme and Module Handbook 2020-2021 (archived)
Module PHIL42415: Ethics and Bias in Data Science
Department: Philosophy
PHIL42415:
Ethics and Bias in Data Science
Type |
Tied |
Level |
4 |
Credits |
15 |
Availability |
Available in 2020/21 |
Module Cap |
None. |
Tied to |
G5K823 |
Tied to |
G5K923 |
Prerequisites
Corequisites
Excluded Combination of Modules
Aims
- To introduce students to contemporary debates on ethical issues and bias resulting from the increasingly widespread application of data analytics, statistical modelling and artificial intelligence in society.
- To introduce students to cutting edge philosophical research on these issues and to examine how to apply this research in practice.
- To provide students with tools that will enable them to apply ethical theories and frameworks to practical problems in Data Science.
- To provide procedures and build expertise for examining, identifying and rectifying the sources of bias that can lead a statistical model to produce inaccurate or unjust results.
- To provide students with the knowledge and skills required to research and write about a specific ethical topic under the guidance of members of staff.
Content
- A sample of the topics covered may include:
- Accountability and transparency in AI.
- Creating trustworthy algorithms.
- Social media and the fracturing of civil society.
- Algorithmic bias
- Unconscious bias
- Gender bias in language modelling.
- Racial bias in facial recognition.
- The use of private medical data for public health purposes.
- Ethical problems with algorithmic approaches to policing, sentencing and probation.
- Data manipulation and collective action.
Learning Outcomes
- Students will be able to:
- Understand the requirements for building ethically robust statistical models.
- Understand the main ethical challenges arising from the use of private data in the public and commercial sphere.
- Understand the societal biases that can affect a statistical model when the algorithms and training data are skewed by prejudice.
- Apply ethical thinking and studies to real-life cases and examples in Data Science.
- Understand the background issues that shape the debate and influence current discussion in the field.
- Be able to draw parallels between different kinds of cases and examples by means of conceptual analysis and philosophical theory.
- Students will be able to:
- Identify key issues, questions and debates regarding the ethics of AI.
- Draw analogies between these issues, questions and debates.
- Identify and make use of relevant literature.
- Identify a philosophical problem, formulate a philosophical position and employ critical skills to address the problem.
- Write an essay which answers a question in an appropriately focused manner, with a clear and concise discussion of the topic area and a structured argument.
- Students will be able to:
- Identify and locate research materials.
- Write in a clear and rigorous style.
- Manage their time efficiently.
- Pursue interdisciplinary research.
- Make a responsible decision about their chosen essay topic.
- Think clearly and independently about the intersection of statistics and society.
Modes of Teaching, Learning and Assessment and how these contribute to
the learning outcomes of the module
- This module will be delivered by the Department of Philosophy with input from Computer Science.
- Most of the teaching will take the form of linked lectures and seminars. Each lecture will last one hour and will be followed by a one-hour seminar, in which there will be a discussion that follows on from the subject of the lecture. In the seminars, we will address questions that are central to the ethics & bias of AI and data analytics and apply ethical thinking to real life cases. Students will have the opportunity to ask questions and debate the topics outlined in the lecture, and will be encouraged to develop their own opinions and defend their own points of view with the help of philosophical concepts and distinctions. They will be guided through the material and have a chance to develop both their analytic and argumentative skills.
- The tutorials will enable smaller groups of students to target a specific research area (based on the essay topic they have chosen) and participate in in-depth discussions of this particular topic. They will have a chance to examine the wider ramifications of their research area and reflect on its practical relevance in Data Science. These tutorials will also enable students to work on their essay-writing techniques, receiving individual guidance where appropriate.
- Towards the end of the module the students will attend a workshop focusing on specific applications of the theories they have studied. During this workshop students will present a team-based case study. They will defend their arguments by responding to questions. This will help students to develop their skills for collaborative ethical decision making.
Teaching Methods and Learning Hours
Activity |
Number |
Frequency |
Duration |
Total/Hours |
|
Lectures |
4 |
Once per week (Term 1, weeks 6-9), before the week’s seminar (they can be timetabled back-to-back) |
1 hour |
4 |
|
Seminars |
4 |
Once per week (Term 1, weeks 6-9), after the week’s lecture (they can be timetabled back-to-back) |
1 hour |
4 |
|
Tutorials |
3 |
Once per week (Term 1, weeks 6-8) |
1 hour |
3 |
|
Workshop |
1 |
Once (Term 1, week 9) |
3 hours |
3 |
|
Preparation and Reading |
|
|
|
136 |
|
Total |
|
|
|
150 |
|
Summative Assessment
Component: Essay |
Component Weighting: 100% |
Element |
Length / duration |
Element Weighting |
Resit Opportunity |
Essay |
3000 words |
100% |
|
One formative preliminary draft of the summative essay (1,500 words). A workshop presentation on the topic of the summative essay.
■ Attendance at all activities marked with this symbol will be monitored. Students who fail to attend these activities, or to complete the summative or formative assessment specified above, will be subject to the procedures defined in the University's General Regulation V, and may be required to leave the University