Part II: Mechanics & Application

Make AI work for you

A 4-Class Module on AI Systems, Mechanisms, and Behaviour


What This Module Is

Most introductions to AI focus on outputs.

This module focuses on process.

We look inside AI systems to understand:

AI systems are not magic. They are pipelines of mechanisms.

Once you see the structure, you can begin to reason about:


The Arc

Class Title What Happens What You Learn
1 Machine Learning 101 We translate real problems into AI workflows Basic lay of the land and key principles of machine learning
2 Neural Network & Generative Modelling We examine how models transform inputs into outputs Neural networks, extensions (images and graphs), and the role of probability & randomness in generative modelling
3 Extending LLMs We examine how LLMs can be augmented into powerful agents, and supported course project development Components and principles of LLM-powered applications

What We’re Asking You to Do

In this module, you are asked to shift perspective.

Not from user → output.

But from system → mechanism → behaviour.

Your task is to:

  1. See the pipeline.
    Understand how an AI system breaks a problem into computational steps.

  2. Locate the mechanism.
    Identify the component that actually produces behaviour.

  3. Explain the behaviour.
    Show how structural design decisions create strengths, limits, and failure patterns.


Learning Outcomes

By the end of this module, you will be able to:

  1. Translate real-world problems into AI system pipelines
  2. Identify and explain mechanisms within AI systems
  3. Analyse how inputs → transformations → outputs produce observable behaviour
  4. Reason about how mechanisms, datasets, and responsibility considerations interact
  5. Communicate system behaviour through clear diagrams and structured explanation

How You’ll Be Assessed

This module is assessed through A1.2 — Case Study 2: The Revealing Mechanism, an individual case study focused on explaining how a specific AI mechanism works.

You will analyse an AI system and show:

Full requirements and deliverables are provided in the assignment-specific outline [link].

This should be read alongside the general case study rubric [link].


Resources


The One-Liner

“If you can’t explain the mechanism, you don’t understand the system — and you can’t make AI work for you.”