Most professionals encountering AI tools today hear four terms used interchangeably: artificial intelligence, machine learning, deep learning, and automation. In casual conversation these words often get treated as synonyms. They are not synonyms. They describe four distinct kinds of systems that work through different mechanisms and have different capabilities, and treating them as interchangeable produces persistent misunderstandings about what any given tool can actually do.
The distinction matters in practical terms. A practitioner who understands what layer of the stack a given tool is operating at can form accurate expectations about its behaviour. A practitioner who does not will routinely be surprised by what the tool can and cannot do, in both directions. They will be surprised when an automation tool fails to handle an exception because it was designed to follow a fixed path, and they will be surprised when a large language model produces a confident-sounding answer that is factually wrong because it was designed to predict plausible text rather than retrieve verified facts. Both failures are predictable from the nature of the underlying system, and professional practice with AI depends on recognising which system is operating.
This module develops the four-layer framework that most of the AI field uses to organise these distinctions, then focuses specifically on what a large language model is as a system, because large language models are the form of AI that most professionals will actually work with day to day. The module assumes no technical background. It works through mechanism explanations in plain language, defines the specific terms the practitioner will encounter (token, context window, inference, hallucination, fine-tuning, retrieval-augmented generation), and connects each concept to the practical consequences for professional work.
Module 1.1 established the historical arc and the transformer architecture at a high level. This module develops the conceptual framework that sits alongside that history. Together, the two modules give the practitioner a working answer to the question of what AI actually is. The modules that follow then develop where it shows up in professional work, how it fails, and how to communicate with it effectively.