Area 00 · 3 lessons

Understand the Machine

Free — no account needed to read

You can only direct what you understand.

In one line: Know enough about how an AI answer is produced to tell a strong one from a confident one.

When you need it

• An assistant hands you a number, a date or a source and you are about to act on it.
• You are deciding whether a task is worth handing over at all.
• Someone states flatly what AI is and is not capable of, and you have no way to check them.

The common misreading

What people think

An AI assistant is basically a search engine that writes.

What is true instead

It generates an answer from patterns and from the context you gave it. Nothing was looked up unless a tool went and looked — so "where did this come from?" is always a fair question.

After this area you can say what an AI answer is actually made of, and which parts of it need checking before you use them.

The key idea

AI systems can do sophisticated reasoning-like work, but how reliable they are depends on the task, the context, the model and the information they were given.

What it teaches

• What modern AI is
• Models, training and inference
• Context and multimodal AI
• Agents, tools and automation
• Where AI is strong
• Where it is currently unreliable

Think first

An AI assistant tells a friend that a medicine is safe to take with their current prescription, and gives a confident explanation.

What would you question before anyone acted on this? Who should actually be asked?

The lessons

• What modern AI actually is — An AI answer is a generated response shaped by training and by the context you give it — not a fact retrieved from a trusted source. (5 min)
• Where AI is genuinely strong — Use AI aggressively for speed, scale, drafts, options and transformation — that frees your time for the parts that need you. (4 min)
• Where AI is currently unreliable — Treat AI output as a capable first draft from a colleague who has never met you: useful, fast, and in need of checking. (6 min)

Practise: The verification drill

Build the habit of checking before trusting.

1. Ask an AI assistant a factual question about something that matters to you (a law, a price, a date, a health fact).
2. Write down its answer and how confident it sounded.
3. Check the answer against two independent sources.
4. Record what was right, what was wrong, and what was missing.

Reflect

What surprised you most about how AI works — and does it change how much you trust it?