1. A conversational engine
You can ask questions in ordinary language, refine the answer, ask for alternatives and challenge assumptions. The conversation itself becomes working context.
Introduction 1
What AI actually is
Modern generative AI is best thought of as a powerful pattern-and-language engine. It can explain, transform, compare, draft, reason through many problems and operate tools when connected to them — but it is not an oracle and it does not automatically know your private context.
Three useful mental models
2. A transformation engine
Give it material and ask for a transformation: summarize, explain, classify, rewrite, compare, extract structure, generate examples or critique an argument.
3. A reasoning worker with tools
When software grants it tools, AI can search, read files, call APIs or perform bounded actions. The tool permissions — not the AI's confidence — define what it can really access.
What it is good at
AI is useful when the task benefits from language understanding, synthesis, iteration or pattern recognition. Typical examples include learning an unfamiliar subject, brainstorming, writing and editing, converting messy notes into structure, reviewing code, producing test cases and exploring several approaches quickly.
Think “accelerator”, not “automatic authority”. A good AI can shorten the path from question to candidate answer enormously. You still decide how much evidence the answer requires.
What it cannot guarantee
An AI may produce a fluent answer that is wrong, outdated, incomplete or based on an assumption you never intended. It can also misunderstand a local abbreviation, private project rule or unstated goal.
Confidence of wording is not confidence of fact. For current facts, important decisions, money, health, law, security or production systems, require appropriate external verification.
Free AI is enough to begin
You do not need a GPU, an API key or programming knowledge to learn useful AI habits. Start with a mainstream hosted assistant and practise the interaction.
Simple first uses
“Explain compound interest without assuming I know finance.”
“Compare these three options. Tell me what information is missing before recommending one.”
“Here are my rough notes. Organize them into a clear one-page briefing.”
Better habit from day one
Tell it what you are trying to achieve, what you already know, what constraints matter and what form the answer should take. Then ask it to identify uncertainty rather than filling gaps silently.
That is the subject of the next page →