Alan Dix

Summary

Social justice is concerned with the overall balance of opportunities, benefit and power within society. It is hard to define precisely, but is broadly about equity, fairness and empowerment, especially protecting the weak or disadvantaged. There are some clear issues such as explicit racial or gender discrimination, or the existence of poverty amongst plenty in many societies. However, there are also more subtle influences that can lead to unfair outcomes for different groups or perpetuation of disadvantages. AI can exacerbate both the more and less obvious forms of social injustice, but can also be an opportunity to expose, challenge or redress inequity.

This course will look at the influence of AI on social justice using two lenses. First using an approach developed during the UK EPSRC funded Not Equal project a few years ago to look at digital technology more generally, which considers three main areas: avoiding harm, doing it right and taking positive action, all dependent on access to technology and broader issues of governance. The second distinguishes ‘what it does’ the direct use of AI and its ethical and social implications vs the way AI potentially changes the very nature of society, even for those not using AI.

The course will include case studies, concrete examples, and opportunities for discussion. It will also include a deep dive into issues of AI bias and explainable AI.

Syllabus

What is social justice?

• End state (e.g. Rawls) vs process vs putative process (Nozik)
• Examples – local discounts?
• Exercises – vineyard, pot of gold
• Ethical conflicts

What AI does

• Avoiding harm
• Doing it right
• Positive action
• Access

Bigger picture – how AI changes society

• Context – social, physical, environmental
• The new AI – scale and power
• Network externalities, positive feedback and emergent monopolies
• Impacts: democracy, inequality, environment
• Signs of hope

Bias, discrimination and explainability

• Different forms of bias
• Sources of bias in AI, proxy variables
• Dealing with bias: visualising bias, de-biasing data
• Definitions of fairness, base rates and fairness conflicts

Case studies / application areas (used at different stages)

• Democracy – when AI ruled France
• Education and examination – UK A’levels in Covid
• Health – diagnosis and outcomes (base rates again!)
• Criminal justice – COMPAS, facial recognition

References

Crivellaro, C. and Dix, A. (2026). AI for Social Justice. CRC Press, Boca Raton, FL. https://alandix.com/ai4sj/ (Although this will not be published at the time of the course, related lecture notes are available from the book web site.)

Dix, A. (2025). Artificial Intelligence – Humans at the Heart of Algorithms, 2nd Edition, Chapman and Hall. https://alandix.com/aibook/ (See especially part IV: Chap. 20 ‘When Things Go Wrong’, Chap. 21 ‘Explainable AI’, and Chap. 23 ‘Philosophical, Ethical and Social Issues’)

Angwin, J., Larson, J., Mattu, S. and Kirchner, L. (2016). Machine bias there’s software used across the country to predict future criminals: And it’s biased against blacks. ProPublica, 23 May 2016. https://www.propublica.org/article/machine-bias-risk-assessments-incriminal-sentencing (Much discussed article exposing dangers of the use of AI in the criminal justice system.)

Larson, J., Mattu, S., Kirchner, L. and Angwin, J. (2016). How we analyzed the COMPAS recidivism algorithm. ProPublica, 23 May 2016. https://www.propublica.org/article/how-weanalyzed-the-compas-recidivism-algorithm (Detailed analysis underlying the Angwin, et al. 2016 article – demonstrates some of the difficulties of defining ‘fairness’.)

Dix, A. (1992). Human issues in the use of pattern recognition techniques. In Neural Networks and Pattern Recognition in Human Computer Interaction, R. Beale and J. Finlay (eds). Ellis Horwood. 429-451. https://alandix.com/academic/papers/neuro92/ (The first paper to warn of the dangers of gender, ethnic and socio-economic bias in black-box machine learning systems.)

Panic, B. and Arthur, P. (2024). AI for peace. CRC Press, Boca Raton, FL. (A rich analysis of the way AI can be used positively to promote peace as well as counter the ways it is being used for the opposite.)

Vallor, S. (2024). The Turing Lectures: Can we live with AI? The Alan Turing Institute. 9 December 2024. (See section@21:27: The Mirror Metaphor for AI) https://www.youtube.com/live/7iX-wiKvYHs?si=W-QSferpGWZFcItI&t=1285

Stone, E. (2024). AI shifts the goalposts of digital inclusion. Joseph Rowntree Foundation, 8 February 2024. https://www.jrf.org.uk/ai-for-public-good/ai-shifts-the-goalposts-of-digital-inclusion

Liebowitz, S., and Margolis, S. (2001). Network effects and the Microsoft case. Chapter 6 in Dynamic competition and public policy: Technology, innovation, and antitrust issues, J. Ellig (ed.), pp.160–192. https://personal.utdallas.edu/~liebowit/netwext/ellig%20paper/ellig.htm

Hartmann, D., Wang, S.M., Pohlmann, L. and Berendt B. (2025). A systematic review of echo chamber research: comparative analysis of conceptualizations, operationalizations, and varying outcomes. J Comput Soc Sc 8, 52. https://doi.org/10.1007/s42001-025-00381-z

Bentley, S.V., Naughtin, C.K., McGrath, M.J., Irons, J.L. and Cooper, P.S. (2024). The digital divide in action: how experiences of digital technology shape future relationships with artificial intelligence. AI Ethics 4, 901–915. https://doi.org/10.1007/s43681-024-00452-3

Pre-requisites

No pre-requisites. The course will describe statistical and algorithmic aspects of AI pertaining to social justice, but will introduce these without assuming prior technical knowledge. Similarly it will dig into social, economic and legal aspects, but only assuming general knowledge in these areas.

Short bio

Alan Dix is Emeritus Professorial Fellow at Cardiff Metropolitan University and Swansea University. He started his academic career as a mathematician and was part of the British team to the International Mathematical Olympiad in 1978. However, he is best known for his work in human–computer interaction (HCI), including writing one of the key textbooks in the area. He was elected to the ACM SIGCHI Academy in 2013 and is a Fellow of the Learned Society of Wales. Outside academia, Alan has been co-founder of two dotcom-era tech companies, developed intelligent lighting and worked in local government and even submarine design.

In every role, Alan seeks to understand and innovate in all aspects where people and technology meet. He has often been prescient in recognising the implications of digital technology, in 1990 writing the first paper on privacy within the HCI literature and in 1992 predicting the potential danger of social, ethnic and gender bias in black-box machine learning algorithms. Alan writes and talks extensively on the connections between artificial intelligence and human issues, both in terms of individual user interfaces and also social implications of technology. This has included leading the algorithmic social justice theme within the UK Not-Equal programme and participating in the European TANGO project on synergistic human–AI decision making. He has also worked on fundamental models of emotion in AI and practical applications of genetic algorithms in engineering.

His latest book, Artificial Intelligence – Humans at the Heart of Algorithms, includes substantial material on social, ethical and human-centred aspects of AI as well as being a broad and accessible introduction to the technical aspects. His other books include Human–Computer Interaction (with Janet Finlay, Gregory Abowd and Russell Beale), one of the key international textbooks in the area; TouchIT: Understanding Design in a Physical-Digital World (with Steve Gill, Devina Ramduny-Ellis and Jo Hare) on the design of physical-digital products; and Statistics for HCI: Making Sense of Quantitative Data. He is also completing two books in the CRC/Taylor & Francis “AI for Everything” series: AI for HCI and AI for Social Justice (with Clara Crivellaro). This course is based on the last of these.