Course Description

This course explores the ethical challenges AI systems present and measures to build them responsibly via the Delft Design for Values methodology. It provides an avenue to implement ethical standards into AI design, ensuring mapped values align with programming and development, focusing heavily on practical scenarios, such as healthcare, under expert guidance from TU Delft professionals.

What Students Will Learn

• Identify and articulate ethical dilemmas in AI development.
• Analyze and forecast potential ethical pitfalls in various AI applications.
• Sketch actionable steps towards embedding ethics in AI utilization.
• Execute responsible AI design protocols.

Prerequisites

While a basic understanding of technical AI development is beneficial, it is not a strict requirement for this intermediate-level course.

Course Coverage

• Understanding AI Ethics and the associated challenges.
• Applying the Design for Values methodology.
• Tools for operationalizing ethics into concrete design requirements.
• Case studies across diverse sectors to illustrate practical application.
• Development of skills in bias identification, transparency, and accountability in AI systems.

Who This Course Is For

This program is tailored for professionals involved in AI system development, alongside managers supervising such projects, aiming to infuse ethical considerations into their workflows.

Real-World Applications

Skills acquired from this course can be used to enhance decision-making in technological design, ensuring AI systems are both effective and ethically conscious, reducing risks and increasing trust in automated processes. Such practices are particularly valuable in sectors where AI impact is critical such as healthcare, government, and industry.

Syllabus

Week 1: Ethical AI Overview

Introduction to the ethical challenges in AI, stakeholder values, and application of Design for Values in healthcare AI.

Week 2: Trust and Explainability

Focus on AI system trustworthiness, accuracy, reliability, and tools for enhancing system explainability.

Week 3: Addressing Bias

Detailed discussion on data bias, algorithm fairness, and methods to monitor and mitigate biases.

Week 4: Accountability and Oversight

Exploration of accountability in AI mishaps and the design of organisational structures for responsible AI use.

Week 5: Handling Value Conflicts

Strategies to manage conflicts between differing ethical values and the final assignment on ethical value design translation.