AI Accountability: Ethical Data Engineering

Course Overview:

This course begins with foundational concepts, defining AI from an EU perspective and linking accountability to AI systems. It delves into the relationship between machine learning and data processing, highlighting the data protection challenges associated with machine learning models. The course explains statistical models and their implications for data protection, emphasizing risk-based approaches to AI and data protection.

Students will gain insights into technical and organizational measures under EU data protection legislation and explore how to engineer ethical AI through data protection by design and by default. The course also covers the development and implementation of AI and ML applications in line with the UK’s data protection framework, detailing the six pillars of corporate AI and data protection strategy.

Course Highlights:

• Comprehensive understanding of AI accountability and ethical data engineering.
• Practical insights into managing AI governance and risk management.
• Step-by-step guidance on establishing a meaningful risk appetite for AI and data protection.
• Exploration of allocative and representational harms in AI systems.
• Techniques for capturing human involvement in automated decision-making processes.
• Detailed discussion on data processing roles and responsibilities in AI technology.
• Methods for assessing necessity, proportionality, and algorithmic impacts of AI systems.
• Emphasis on the importance of regular algorithmic impact assessments and adapting to concept drift.
• Case studies on data processing relationships in different AI contexts, including cloud services and financial services.