As technology evolves, the associated ethical challenges become more and more multi-faceted and complicated. It is vital that students have a clear understanding of some of these challenges that they will face in a professional context, and that they are given tools to aid them to deal with these challenges. In this module it is not envisioned that the lecturer will set out to tell the students the “correct” answers to these issues, but rather that through discussion and reflection the students come to develop a set of principles for themselves that are not in conflict with relevant legislation and professional guidelines.
In conjunction with the philosophical elements of this module, there will be a very strong practical element to this module, where the students will be reviewing, amending, collating and (potenitally) creating datasets, as well as examining and amending (including the tuning the hyperparameters of) alogrithms that may be exhibiting potential ethical concerns. When considering datasets some key ethical considerations that the students must reflect on include: bias, privacy, the reflection of reality in datasets, the correlation-causation dilemma, and the alignment of practice with professional principles.
A number of professional bodies (including the Association for Computing Machinery, the British Computer Society, and the Institute of Electrical and Electronics Engineers) have developed a series of ethical guidelines and Code of Conducts that will be examined in this module, however the reality of the situation is that as technology evolves (and is deployed in novel ways) some existing ethical guidelines will evolve.
Weeks 1-2: Ethical Principles
- AI Ethical Themes: Safety, Transparency, Trustwothiness, Explainability, Beneficence, Non-maleficence, Autonomy, Justice, Explicability
- Usability Ethics: Accessibility, Universal Design, Globalisation, Internationalisation, and Localisation ethical challenges, Technophobia.
Weeks 3-4: Ethical Paradigms
- The Principles of Respect for autonomy, Beneficence, nonmaleficence, and justice. Introducing Ethical philosophers
- Ethical Frameworks: Software Engineering Code of Ethics and Professional Practice (ACM/IEEE-CS), Computer Ethics Institute (CEI) 10 commandments, BCS Codes of Conduct and Practice, Electronic Frontier Foundation, EU Trustworthy framework, ALTAI.
Weeks 5-6: Ethical Datasets
- AI Ethical Considerations: Bias, Privacy, Reflection of Reality, Correlation-Causation, and Alignment with Professional Principles
- Anonymising datasets, k-anonymity, l-diversity, Differential privacy, de-anonymization. Ethical Hacking: Cybercrime, Legal responsibility, computer security, technological and tools
Weeks 7-8: Bias in Datasets
- Bias in Data (racism, sexism, etc.), Confidence in Data (Dataset size), Visualisation biasing, Statistical biasing, unacknowledged data collection (GPS tracking, microphone, and camera activation without the user’s consent), GDPR and Data Protection legislation.
- Algorithmic Ethics: Bias in algorithms (racism, sexism, etc.), Lack of explainability of some algorithms, value-based development, Software
Weeks 9-10: Literature and Legislation
- Processes and procedures for Ethical Approval, Informed Consent: Ethical Research, Terms and Conditions, Right to withdraw, anonymity, Right to be forgotten. Intellectual Property: Copyright, Creative Commons, Software Patents, Copyright on the WWW, Open Source software.
- Student Ethics: Plagiarism, email and Social Media use and abuse, Integrity, Confidentiality, Accountability, Conflicts of Interest, Fundraising.
Weeks 11-12: Technology Solutions
- Testing is ethical imperative, computer security is ethical imperative, Selective Censorship of WWW content. Personalisation of WWW content. Robot Ethics: Driverless Cars, Drones, Internet of Things, Home Assistants (Suri, etc.).
- Case Studies in topics such as:
- Driverless car fatalities
- Computer Security failures
- Computer Testing failures
- Green IT
- Doxxing
Week 13: Exams Review
- Exams and Study Technique
- Topics Reviews
This module will employ teaching methods and learning situations in the traditional roles such as lectures, seminars, tutorials and labwork, as well as more innovative, Student-based learning methods such as problem solving in groups for both theoretical and practical situations.
Central to this programme and this module are the detection and elimination of potential ethical concerns in AI. In this module that will be specifically manifested in the use of datasets to teach issues such as bias, privacy, reflection of reality, correlation-causation, and alignment with professional principles. Therefore the students will be required to check the characteristics of the data (mean, mode, median, standard deviation, skewness, kurtosis, etc.), and seriously consider (and justify) their treatment of both missing data and outliers. It will be insufficient to simply remove atypical data, the students must learn to investigate, supplement and justify their decisions in their treatment of this data, and they will have to do this practically with real datasets. In some cases studentgroups may end up creating their own datasets, or significantly augmenting existing datasets. Additionally an exploration of ethical issues in algorithms (bias, privacy, reflection of reality, correlation-causation, and alignment with professional principles) will be explored using practical examples also. Where appropriate, students will provide feedback from group research through cascading the knowledge to peers and through presentations.
The prevalence of technology in some countries means that an enormous amount of data is being collected and collated on a daily basis, and the use of this data gives rise to significant ethical issues, including privacy (which incorporates considerations such as confidentiality, anonymity, and security). Data from discrete sources can be brought together to produce an extremely detailed picture of the private lives of individuals, therefore Software Developers and Data Scientists have an obligation to ensure that only the data necessary for a given task is made visible to end-users. They are also obliged to ensure that the private data of individuals is protected sufficiently by strong security protocols. Computer professionals face a number of unique challenges in a working context, including corporate governance, human resource challenges, professional boundaries, and professional responsibilities.
Students will be expected to review existing ethical guidelines and Code of Conducts to explore areas where the evolution of technology has meant that some existing ethical guidelines are either out of date, or need to be recontextualized.
Students will be encouraged to be pro-active in their approach to learning through the use of case studies and simulation exercises, working independently and in groups. In some cases Students will be expected to use computer-based learning material to supplement studies.
Guest lecturers from industry and academia will be invited where appropriate to expose students to how topics covered in this module are used within the broader area of ethics.
| Module Content & Assessment | |
|---|---|
| Assessment Breakdown | % |
| Other Assessment(s) | 100 |