Skip down to main content
  • OII >  
  • Study >  

    • Fairness, Accountability, and Transparency in Machine Learning

    Key Information

    Course details
    MSc Option course, Hilary Term
    Assessment
    Coursework submission
    Reading list
    View now
    Tutor
    Dr Ana Valdivia

    About

    Automated decision-making systems are increasingly prevalent across various domains, tasked with filtering, sorting, classifying, recommending, and shaping human decisions and experiences. While these algorithmic solutions present significant opportunities, they also pose substantial risks that require careful examination to avoid unintended and potentially harmful outcomes. Real-world examples have demonstrated that even algorithms designed with good intentions can lead to negative consequences such as racial discrimination.
    This highlights the urgent societal need to educate the next generation of digital practitioners in transdisciplinary methods, fostering a critical understanding of machine learning. This course is designed to explore how algorithmic systems can contribute to social progress while also perpetuating or exacerbating oppression and social injustice. By drawing on critical data studies literature at the intersection of computer science and social science, the course aims to equip students with both computational and qualitative methods necessary to assess and address the impact of algorithmic systems in real-world contexts. The course will focus on discussing key reading from leading conferences such as the ACM Fairness, Accountability and Transparency (ACM FAccT) or the AAAI/ACM Conference on AI, Ethics, and Society, featuring work from scholars across both the Global North and South. It will offer critical and empirical analyses of surveillance technologies, large language models (LLMs), and the environmental impacts of AI. Students are expected to engage critically, reflecting on their own experiences and the course materials to develop a nuanced understanding of the societal implications of algorithmic systems.

    Topics

    • Algorithmic Governance: Human Rights in the Automated and Digitalised Society.
    • Automated Decision-Making systems in the Welfare State, Predictive Policing and Court Systems.
    • Algorithmic Fairness: Conceptualisation, Definitions and Metrics.
    • Explainability and Interpretability Methods. Causality and Graphical Discrimination Analysis.
    • Algorithmic Accountability: Definitions, Frameworks and Limitations.
    • Surveillance Technologies, Ethics and Fairness.
    • Natural Language Processing, Ethics and Fairness.
    • Environmental Impact of AI Technologies.

    Learning Outcomes

    At the end of this course, you will have developed skills for:
    • Understanding how fairness is conceptualised in a range disciplines, from social science to computer science, as well as in many practical applications of machine learning.
    • Critically interrogating how foundational aspects of existing and emerging technologies, policies and practices connect to unintended harms, and effectively identified algorithmic benefits, risks and limitations.
    • Becoming acquainted with computational and qualitative methods to effectively build and implement audit and evaluation practices on data and algorithmic systems following current legal frameworks such as the AI Act.
    • Understanding the theoretical and empirical global environmental, social, and legal impacts of decision-making systems from different perspectives and on different communities both in the Global North and South.

    Course Tutor