Ethics and Trustworthy AI

This part is being prepared. The notes will cover how to use AI responsibly: who is accountable when systems cause harm, how to think about fairness and privacy, and what the law expects of high-risk applications.

The planned sections follow the course table of contents.

Accountability, fairness, privacy, and environmental impact

AI systems are used in decisions that affect people: hiring, credit, policing, healthcare, and more. This section will look at accountability (who is responsible when something goes wrong), basic fairness, privacy, and the environmental cost of training and running large models. The Moral Machine experiment is one well-known example of how hard it is to encode values into autonomous systems.

Explainable AI and causal modelling

Many modern models are hard to interpret. This section will introduce explainable AI: why explanations matter, what kinds of explanations are useful, and how causal modelling helps separate correlation from the effects we actually care about.

Formal treatment of AI for Science

AI is increasingly used in scientific discovery. This section will discuss, at a conceptual level, what it means to use AI as a scientific tool, and what standards of evidence and reproducibility that implies.

EU AI Act

The EU AI Act is a risk-based legal framework for placing AI systems on the market in the European Union. This section will outline the main idea of the Act and what it means for high-risk systems, without attempting a full legal commentary.