CSC-30045 - Module Specification
School of Computing and Mathematics
Faculty of Natural Sciences

Programme/Approved Electives for 2023/24

Available as a Free Standing Elective

Co-requisites

None

Prerequisites

None

Barred Combinations

None

Description for 2023/24

The rise of AI and Data Science and AI within society has been filled with potential but has also come with increasing risk and harms. This includes risks to personal privacy and freedoms, impact of discriminatory algorithms due to algorithmic bias and online harms. As a result we see a mixture of over-confidence in the value and accuracy of AI systems resulting in tech solutionism in juxtaposition to increasing criticism and lack of public trust. In response to this the need for ethically aligned design of AI and data science projects has grown. This module will outline the risks and harms by utilising a case study approach and will explore the body of regulation, governance frameworks and techniques. This will provide the students the opportunity to develop the skills and knowledge required by employers for ethically aware data scientists.

This module will provide the students with an understanding of Data Ethics and Security. They will explore relevant regulations, governance frameworks and standards required for ethically aligned design and how these relates to different aspects of the data science lifecycle. They will apply this knowledge via techniques and tools that can be applied to areas such as data anonymisation, debiasing and fairness testing.

Intended Learning Outcomes

Debate general ethical concepts, such as deontological and teleological ethics, and consider how they can be applied to data science and valuate how AI relates to aspects of Human Rights and UN Sustainable Development Goals: 1
Apply the different methods for data anonymisation and psuedoanonymisation and how to protect identification: 1
Analyse datasets and algorithmic outcomes for bias by using summative statistics, performance metrics and fairness tests. This will include demonstrating an understanding of protected characteristics and proxy features and methods for debiasing and/or mitigating for bias: 1
Compare the data security techniques that can be employed to ensure data privacy and security: 1
Test the legislative and governance landscape associated with ethical data science including key legislation such as GDPR, Data Protection Act and international standards and ethics panels: 1
Debate the role of Explainable AI (XAI) methods and techniques in their provision of Data Science and AI solution that can be understood by humans: 1

Study hours

10 hours of practical activities
20 hours of lectures
120 hours of independent study

School Rules

None

Description of Module Assessment