AI is in every headline. Start with what it actually is.

You have already cleared up one of the biggest confusions in AI: that it is not one thing, but a family of related technologies, each with its own methods and limits. Now comes the question most people do not know they can ask: how does any of this actually work? Not in a hand-wavy sense of 'it learns from data', but in a real sense. What is the model doing when it learns? What happens inside the system when you type a message and press send? Why does it sometimes produce confident, fluent nonsense, and why can it not remember your conversation from yesterday? These are not trick questions. They have clear, non-mathematical answers, and understanding them changes how you use AI tools, how you evaluate claims about them, and how you think about their limits. Here is the central insight this lesson builds towards: modern AI models are not databases that look things up, not rule-following systems executing step-by-step instructions, and not reasoning engines in any philosophical sense. They are pattern-extraction machines, trained to recognise statistical regularities in enormous quantities of data and to use those patterns to generate plausible-seeming outputs. That might sound deflationary, but it is the opposite. The fact that this approach produces systems that can write code, translate languages, analyse images, and hold coherent conversations is genuinely remarkable. This lesson covers the full chain: from the basic logic of machine learning, through the structure of neural networks, to the architectural innovation that made modern large language models possible, and finally to what actually happens when you use one. By the end, you will have a mental model that holds up under scrutiny.

In February 2024, a finance employee at the Hong Kong office of a respected global engineering firm joined a video call with several colleagues, including what appeared to be the company's chief financial officer. The conversation seemed normal, the instructions were clear, and the employee authorised a transfer of around 25 million US dollars. Every person on that call, except the employee making the transfer, was a deepfake. This is no longer the world in which robotic-sounding text, wonky hands, or pixelated faces would give the game away. AI-generated content, whether text, images, audio, or video, has reached a quality threshold where casual inspection frequently fails, and the question of whether something is real or synthetic has become one of the defining media literacy challenges of the decade. What makes this harder is that the tools most people reach for first, automated AI detectors, are substantially less reliable than advertised. By the end of this lesson you will understand the four types of synthetic content and how they are produced, why automated detectors are unreliable, what to actually look for in each modality, how provenance systems work, and how to develop the right kind of scepticism: not paranoid, not credulous, but calibrated.

You have opened an AI chat tool. The cursor blinks, and you type something like write me an email or explain this to me. A response appears: technically coherent, rather generic, and not quite what you wanted. This is the most common experience people have when they first use a large language model: the AI is capable of remarkable things, but what it actually produces seems oddly flat. The usual conclusion is that the tool is overhyped. The more accurate conclusion is that the brief was thin. An AI language model works only from what you give it. There is no background knowledge about you, your organisation, your audience, or your intentions unless you supply it. Give it something vague and it will return something vague, confidently and fluently, and it will not tell you it did not have enough to go on. The good news is that effective prompting is not a technical skill, it is a communication skill: the same principles that make a good brief, a clear email, or a precise question also make a good prompt. This lesson gives you a practical framework for applying those principles, along with worked examples you can adapt immediately.

You have sat in the meeting. Someone mentions a large language model, someone else talks about fine-tuning, and before long the conversation has drifted into inference pipelines and responsible AI frameworks. You nodded. Everyone nodded. But if you are honest, at least three of those phrases could have meant almost anything. You are not alone, and you are not behind. The vocabulary around AI has expanded faster than almost any technology field in recent memory, and even people who use these tools every day frequently misuse the terms. This lesson cuts through that. We have chosen ten terms you will genuinely encounter this week, whether you work in finance, education, healthcare, or public services. For each one you will get a clear definition, the most common misconception to watch out for, and where relevant a concrete analogy to make it stick. By the end you will be able to follow technical conversations with confidence, ask better questions of AI vendors and colleagues, and spot when jargon is being used to clarify rather than to obscure.