Postgraduate Course: Algorithmic Bias, Fairness and Justice (Online) (EFIE11467)
Course Outline
| School | Edinburgh Futures Institute |
College | College of Arts, Humanities and Social Sciences |
| Credit level (Normal year taken) | SCQF Level 11 (Postgraduate) |
| Course type | Online Distance Learning |
Availability | Available to all students |
| SCQF Credits | 20 |
ECTS Credits | 10 |
| Summary | This course introduces students to one of the most important challenges in academia and industry for making data-driven AI systems ethical: algorithmic bias. By drawing on interdisciplinary perspectives from computing, statistics, political and legal philosophy, and feminist and media theory, students will learn about principles for addressing fairness in AI and machine learning systems, how they relate to the limits and opportunities of algorithmic fairness in with regard to the aims of social and distributive justice. |
| Course description |
Data-driven machine learning systems learn from biases in existing data and risk reproducing and amplifying social patterns of bias reflected in this data, such as racial or gender bias. They also rely on the consumption of large amounts of energy and fresh water, contributing to climate change. Such systems then exacerbate existing injustices. Algorithmic fairness is an emerging research area for addressing societal biases in the hope of making data-driven AI systems ethical. This course introduces students to issues of bias, fairness, and justice in data-driven machine learning systems by drawing on a diverse set of perspectives from computing, statistics, political and legal philosophy, and feminist and media theory in order to raise critical awareness and understanding of these issues as well as providing students with tools and knowledge to examine and consider such problems.
Topics include sources of bias in data and machine learning, methods for measuring and mitigating bias and unfairness, notions of individual and group fairness and the tensions between them, limitations of algorithmic fairness approaches, environmental impacts of machine learning systems, and philosophical accounts of distributive and structural justice relevant to machine learning systems. You will work together in collaborative groups to practice the identification and evaluation of algorithmic bias concerns in concrete cases; students will also practice jointly deliberating about and communicating the benefits and limits of different methods, techniques and approaches to algorithmic fairness and justice.
Edinburgh Futures Institute (EFI) - Online Hybrid Course Delivery Information:
The Edinburgh Futures Institute will teach this course in a way that enables online and on-campus students to study together. To enable this, the course will use technologies to record and live-stream student and staff participation during their teaching and learning activities. Students should note that their interactions may be recorded and live-streamed (see the Lecture Recording and Virtual Classroom policies for more details). There will, however, be options to control whether or not your video and audio are enabled.
You will need access to a personal computing device for this course. Most activities will take place in a web browser, unless otherwise stated. We recommend using a device with a screen, a physical keyboard, and internet access.
|
Entry Requirements (not applicable to Visiting Students)
| Pre-requisites |
|
Co-requisites | |
| Prohibited Combinations | |
Other requirements | None |
Information for Visiting Students
| Pre-requisites | None |
| High Demand Course? |
Yes |
Course Delivery Information
|
| Academic year 2026/27, Available to all students (SV1)
|
Quota: 0 |
| Course Start |
Semester 2 |
Timetable |
Timetable |
| Learning and Teaching activities (Further Info) |
Total Hours:
200
(
Lecture Hours 20,
Programme Level Learning and Teaching Hours 4,
Directed Learning and Independent Learning Hours
176 )
|
| Assessment (Further Info) |
Written Exam
0 %,
Coursework
100 %,
Practical Exam
0 %
|
| Additional Information (Assessment) |
The course will be assessed by means of the following components:
1) Mock Ethics Statement for Published Paper or Publicly Available AI System (20%)
Students will be divided into groups to write a 750-word mock 'authors ethics statement' for an AI research paper that presents significant fairness and bias concerns.
2) Mock Ethics Panel for Mock Assessment of Mock Ethics Statement (20%)
In this assessment, students will be divided into groups to function as Mock Ethics Assessment Panels evaluating another group's Ethics Statement. Groups will write a 750-word report, detailing their assessment of another group's ethics statement.
3) Fairness Risk Report (60%)
Students will also produce an individual 2,500-word 'fairness risk report' on a selected case study involving an algorithmic model about which there are evident fairness and bias concerns. |
| Feedback |
Feedback on any formative assessment may be provided in various formats, for example, to include written, oral, video, face-to-face, whole class, or individual. The Course Organiser will decide which format is most appropriate in relation to the nature of the assessment.
Feedback on both formative and summative in-course assessed work will be provided in time to be of use in subsequent assessments within the course.
Feedback on the summative assessment(s) will be provided in written form via Learn, the University of Edinburgh's Virtual Learning Environment (VLE).
Formative Feedback Opportunity:
Formative feedback is ongoing feedback which monitors learning and is intended to improve performance in the same course, in future courses, and also beyond study.
Formative feedback will be provided in the immersive phase for the asynchronous groups and to individuals in the Q&A session, when the Course Organiser will jointly help to shape the understandings of students of the core issues and the first collaborative task.
The Course Organiser will be available to meet with groups to discuss work in progress and mediate any significant problems or disagreements within the group that cannot be resolved internally. |
| No Exam Information |
Learning Outcomes
On completion of this course, the student will be able to:
- Demonstrate a basic understanding of key concepts, theories, metrics, and principles of bias, fairness, and justice from statistics, machine learning, feminist, legal, and political philosophy.
- Critically discuss and evaluate a variety of perspectives in debates on how various conceptions of bias, fairness, and justice are to be used for the design of ethical data and AI ecosystems.
- Work constructively with others to weigh bias, fairness and justice considerations and identify potential remedies and interventions for a concrete instance of a data-driven machine learning model.
- Produce and clearly communicate for non-specialists a basic analysis and advisory output pertaining to bias, fairness, and justice in a concrete data-driven AI application.
- Identify and critically evaluate the technical and moral trade-offs involved in decisions about which fairness metrics or interventions to employ in a given AI application context, while weighing these against the broader aims of justice.
|
Reading List
Required Reading:
Angwin, J., & Larson, J. (2016, December 30). Bias in Criminal Risk Scores Is Mathematically Inevitable, Researchers Say. ProPublica. https://www.propublica.org/article/bias-in-criminal-risk-scores-is-mathematically-inevitable-researchers-say
Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016, May 23). Machine Bias: There's Software Used Across the Country to Predict Future Criminals. And It's Biased against Blacks. ProPublica. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
Barocas, S., Hardt, M., & Narayanan, A. (Chapter 1, 2023). Fairness and machine learning: Limitations and opportunities. MIT Press. https://fairmlbook.org
Benjamin, R. (Chapter 1, 2019). Race after technology: Abolitionist tools for the new Jim code. Polity.
Blodgett, S. L., Barocas, S., Daumé III, H., & Wallach, H. (2020). Language (Technology) is Power: A Critical Survey of "Bias" in NLP. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 5454-5476. https://doi.org/10.18653/v1/2020.acl-main.485
Hardt, M. (2016, July 12). How big data is unfair. Medium. https://medium.com/@mrtz/how-big-data-is-unfair-9aa544d739de
Jordan, M. I., & Mitchell, T. M. (2015). Machine learning: Trends, perspectives, and prospects. Science, 349(6245), 255-260. https://doi.org/10.1126/science.aaa8415
Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2022). A Survey on Bias and Fairness in Machine Learning. ACM Computing Surveys, 54(6), 1-35. https://doi.org/10.1145/3457607
Stark, L. (2018). Algorithmic psychometrics and the scalable subject. Social Studies of Science, 48(2), 204-231. https://doi.org/10.1177/0306312718772094
Wallach, H., Desai, M., Pangakis, N., Cooper, A. F., Wang, A., Borocas, S., Chouldechova, A., Atalla, C., Blodgett, S. L., Corvi, E., Dow, P. A., Garcia-Gathright, J., Olteanu, A., Reed, S., Sheng, E., Vann, D., Wortman Vaughan, J., Vogel, M., Washington, H., Jacobs, A. Z. (2024). Evaluating Generative AI Systems is a Social Science Measurement Challenge. Workshop on Evaluating Evaluations (EvalEval). https://doi.org/10.48550/arXiv.2411.10939
Recommended Reading:
Binns, R. (2018). Fairness in machine learning: Lessons from political philosophy. In S. A. Friedler & C. Wilson (Eds.), Proceedings of the 1st conference on fairness, accountability and transparency (Vol. 81, pp. 149-159). PMLR. https://proceedings.mlr.press/v81/binns18a.html
Birhane, A. (2021). Algorithmic injustice: A relational ethics approach. Patterns, 2(2), 100205. https://doi.org/10.1016/j.patter.2021.100205
Blodgett, S. L., Lopez, G., Olteanu, A., Sim, R., & Wallach, H. (2021). Stereotyping Norwegian Salmon: An Inventory of Pitfalls in Fairness Benchmark Datasets. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), 1004-1015. https://doi.org/10.18653/v1/2021.acl-long.81
Buolamwini, J., & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. In S. A. Friedler & C. Wilson (Eds.), Proceedings of the 1st Conference on Fairness, Accountability and Transparency (Vol. 81, pp. 77-91). PMLR. http://proceedings.mlr.press/v81/buolamwini18a.html
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334), 183-186. https://doi.org/10.1126/science.aal4230
Fazelpour, & Danks. (2021). Algorithmic bias: Senses, sources, solutions. Philosophy Compass, 16(8). https://doi.org/10.1111/phc3.12760
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé III, H., & Crawford, K. (2018). Datasheets for Datasets. arXiv:1803.09010 [Cs]. http://arxiv.org/abs/1803.09010
Hao, K., & Stray, J. (2019). Can you make AI fairer than a judge? Play our courtroom algorithm game. MIT Technology Review. https://www.technologyreview.com/2019/10/17/75285/ai-fairer-than-judge-criminal-risk-assessment-algorithm/
Hoffmann, A. L. (2019). Where fairness fails: Data, algorithms, and the limits of antidiscrimination discourse. Information, Communication & Society, 22(7), 900-915. https://doi.org/10.1080/1369118X.2019.1573912
Kasirzadeh, A. (2022). Algorithmic Fairness and Structural Injustice: Insights from Feminist Political Philosophy. Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society, 349-356. https://doi.org/10.1145/3514094.3534188
Kulynych, B., Overdorf, R., Troncoso, C., & Gürses, S. (2020). POTs: Protective optimization technologies. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, 177-188. https://doi.org/10.1145/3351095.3372853
Ledford, H. (2019). Millions of black people affected by racial bias in health-care algorithms. Nature, 574(7780), 608-609. https://doi.org/10.1038/d41586-019-03228-6
Longpre, S., Kapoor, S., Klyman, K., Ramaswami, A., Bommasani, R., Blili-Hamelin, B., Huang, Y., Skowron, A., Yong, Z.-X., Kotha, S., Zeng, Y., Shi, W., Yang, X., Southen, R., Robey, A., Chao, P., Yang, D., Jia, R., Kang, D., ... Henderson, P. (2024). A Safe Harbor for AI Evaluation and Red Teaming (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2403.04893
Mumford, L. (1964). Authoritarian and Democratic Technics. Technology and Culture, 5(1), 1.
https://doi.org/10.2307/3101118
Raji, I. D., Bender, E. M., Paullada, A., Denton, E., & Hanna, A. (2021). AI and the Everything in the Whole Wide World Benchmark. Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks. Conference on Neural Information Processing Systems. https://datasets-benchmarks-proceedings.neurips.cc/paper_files/paper/2021/file/084b6fbb10729ed4da8c3d3f5a3ae7c9-Paper-round2.pdf
Resnik, P. (2025). Large Language Models Are Biased Because They Are Large Language Models. Computational Linguistics. 51 (3). https://doi.org/10.1162/coli_a_00558
Sambasivan, N., Arnesen, E., Hutchinson, B., Doshi, T., & Prabhakaran, V. (2021). Re-imagining Algorithmic Fairness in India and Beyond. Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 315-328. https://doi.org/10.1145/3442188.3445896
Solaiman, I., Talat, Z., Agnew, W., Ahmad, L., Baker, D., Blodgett, S. L., Chen, C., III, H. D., Dodge, J., Duan, I., Evans, E., Friedrich, F., Ghosh, A., Gohar, U., Hooker, S., Jernite, Y., Kalluri, R., Lusoli, A., Leidinger, A., .... Subramonian, A. (2024). Evaluating the Social Impact of Generative AI Systems in Systems and Society (No. arXiv:2306.05949). arXiv. http://arxiv.org/abs/2306.05949
Winner, L. (1980). Do Artifacts Have Politics? Daedalus, 109(1). http://www.jstor.org/stable/20024652.
Further Reading:
Barocas, S., Hardt, M., & Narayanan, A. (2023). Fairness and machine learning: Limitations and opportunities. MIT Press. https://fairmlbook.org
Bender, E. M., & Friedman, B. (2018). Data Statements for Natural Language Processing: Toward Mitigating System Bias and Enabling Better Science. Transactions of the Association for Computational Linguistics, 6, 587-604. https://doi.org/10.1162/tacl_a_00041
Benjamin, R. (2019). Race after technology: Abolitionist tools for the new Jim code. Polity.
Birch, K. (2025). Do Artifacts Have Political Economy? Science, Technology, & Human Values, 01622439251352167. https://doi.org/10.1177/01622439251352167
Joerges, B. (1999). Do Politics Have Artefacts? Social Studies of Science, 29(3), 411-431. https://doi.org/10.1177/030631299029003004
Joerges, B. (1999). Scams Cannot Be Busted: Reply to Woolgar & Cooper. Social Studies of Science, 29(3), 450-457. https://doi.org/10.1177/030631299029003006
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019). Model Cards for Model Reporting. Proceedings of the Conference on Fairness, Accountability, and Transparency, 220-229. https://doi.org/10.1145/3287560.3287596
Woolgar, S., & Cooper, G. (1999). Do Artefacts Have Ambivalence: Moses, Bridges, Winner's Bridges and other Urban Legends in S&TS. Social Studies of Science, 29(3), 433-449. https://doi.org/10.1177/030631299029003005
Xiang, A., & Raji, I. D. (2019). On the Legal Compatibility of Fairness Definitions (No. arXiv:1912.00761). arXiv. https://doi.org/10.48550/arXiv.1912.00761 |
Additional Information
| Graduate Attributes and Skills |
Not entered |
| Keywords | Algorithmic Bias,Algorithmic Fairness,Algorithmic Justice,Fair Machine Learning |
Contacts
| Course organiser | Dr Zeerak Talat
Tel:
Email: ztalat@ed.ac.uk |
Course secretary | Miss Yasmine Lewis
Tel:
Email: yasmine.lewis@ed.ac.uk |
|
|
University Homepage