Group Coursework

1 - Set exercises - 100

To view information for a specific offering, select a value from the drop down above.

<ul><li>Foundations of AI Ethics: Historical context of algorithmic harm; ethical frameworks (deontological, consequentialist, virtue ethics) applied to AI; stakeholder analysis and value tensions; corporate AI ethics principles (Microsoft, Google, IBM); the business case for responsible AI.</li><li>Algorithmic Bias and Fairness: Sources of bias (data, algorithmic, human-in-the-loop); fairness definitions and their trade-offs (demographic parity, equalised odds, calibration); measuring bias in classification and ranking systems; intersectionality and compounding disadvantage; case studies (COMPAS, Amazon hiring, healthcare algorithms).</li><li>Bias Mitigation Techniques: Pre-processing methods (resampling, reweighting, representation learning); in-processing constraints (fairness regularisation, adversarial debiasing); post-processing calibration; fairness toolkits (Fairlearn, AI Fairness 360, What-If Tool); validation and monitoring for fairness drift.</li><li>Explainable AI (XAI): The interpretability-accuracy trade-off; model-agnostic methods (LIME, SHAP, anchors); attention-based explanations; counterfactual explanations; global vs. local interpretability; designing explanations for different stakeholders (data scientists, business users, affected individuals).</li><li>Regulatory Landscape and Compliance: EU AI Act risk classification and requirements; GDPR automated decision-making provisions (Article 22); UK algorithmic transparency standards/ICO guidance; US approaches (FTC, EEOC, Executive Order on AI); Canadian AIDA draft; Singapore AI Verify/PDPA and emerging Asia-Pacific standards; sector-specific requirements (financial services, healthcare); algorithmic impact assessments; documentation and audit trails.</li><li>AI Governance and Risk Management: Organisational AI governance structures; model risk management frameworks (SR 11-7); AI ethics boards and review processes; vendor assessment for AI procurement; incident response and remediation; insurance and liability considerations; building responsible AI culture.</li></ul>

Feedback is provided in a number of ways throughout your module. This includes feedback you receive during teaching from both peers and lecturers and that provided for formative assessment. Following an assessment, you will be given your marks and feedback in line with the assessment regulations and policy. More information on the timing and type of feedback that will be provided for each assessment will be available from the module leader. <a href='https://www.citystgeorges.ac.uk/about/governance/policies/student-policies-and-regulations'>https://www.citystgeorges.ac.uk/about/governance/policies/student-policies-and-regulations</a>

This module will support the development of all the graduate attributes. Your module leader will provide you with information about how your learning, teaching and assessment activities support your development of these attributes. The graduate attributes are: - Always learning - Engaged in the world - Technical and digital - Connected Professional - Creating Impact

Directed Asynchronous - 

In-Person - 

Self-directed - 

 Interpretable Machine Learning: A Guide for Making Black Box Models Explainable 

Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence

Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy

Fairness and Abstraction in Sociotechnical Systems. 

EU Artificial Intelligence Act.

 Mastering AI Governance A Guide to Building Trustworthy and Transparent AI Systems

Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing. 

Guidance on AI and Data Protection

Fairness and Machine Learning: Limitations and Opportunities. 

BBM162 - Responsible AI Deployment

BBM162-PRD3-PER-ALV1

The module combines theoretical foundations in ethics and law with hands-on technical labs using fairness and explainability toolkits. Case studies drawn from high-profile algorithmic failures provide context for understanding real-world stakes. You will engage with primary regulatory texts and guidance documents, developing practical compliance skills. Guest speakers from regulatory bodies, civil society organizations, and corporate AI ethics teams provide diverse practitioner perspectives. Structured debates on contested topics develop your capacity to navigate ethical complexity.

Contact hours are designed to be flexible and are delivered using a variety of methods to help you learn in the most effective way. A proportion of the core content may be delivered online and additional support materials and content will also be made available through Moodle.

Responsible AI Deployment

Conduct algorithmic audits to detect and quantify bias in machine learning models.

Skills

Prepare algorithmic impact assessments and regulatory compliance documentation.

Articulate the regulatory requirements for AI systems under GDPR, EU AI Act, UK ICO guidance, US FTC/EEOC expectations, Canadian AIDA draft, Singapore AI Verify/PDPA, and sector-specific regulations.

Knowledge

Implement bias mitigation techniques using industry-standard fairness toolkits.

Design AI governance frameworks appropriate to organizational context and risk profile.

Critically assess different fairness definitions, their mathematical properties, and contexts where each is appropriate.  Reflecting on the need for contextual judgement between competing definitions.

Explain major AI ethics frameworks and articulate their implications for business AI deployment.

Evaluate third-party AI systems for responsible AI alignment during procurement.

Apply explainable AI methods to generate interpretable outputs for diverse stakeholder audiences.

Assess the limits of technical solutions and the importance of organizational and societal interventions.

Values and attitudes

Actively centre affected communities in AI system design and evaluation.