What you'll be able to do
- Explain, in plain language, what today's AI actually is — and isn't.
- Describe what generative AI does, and why its answers sound so confident.
- Recognise the tasks AI is genuinely good at, and where it fails.
- Decide when it's sensible to reach for AI, and when to be careful.
1. Why this matters
The people around you are already using AI tools — to draft emails, summarise documents, and get unstuck on a blank page. Whether or not you've tried them, AI is quietly becoming part of everyday work.
Here's the important part: the difference between people who get real value from these tools and people who get burned by them is not technical skill. It's understanding what the tool actually is underneath. Get that right, and everything else — writing good prompts, checking output, using it safely — follows naturally.
2. What "AI" means today
Forget the science-fiction version — the conscious robot with its own goals. That doesn't exist. The AI you'll use at work is narrow: very capable at specific language and pattern tasks, but with no awareness, no intentions, and no understanding of the world the way you have.
The kind you'll meet most is generative AI: tools that generate new text (and sometimes images or code) in response to a request you type. They all work on the same basic principle — and it's the one idea worth really understanding.
3. The one idea that explains everything
The core idea
A generative AI predicts the most likely next words, based on patterns in the vast amount of text it was trained on.
It is, in effect, an extraordinarily sophisticated version of the auto-complete on your phone. That single fact explains almost everything about how these tools behave:
- It produces language that is plausible, not language that is verified. It's optimised to sound right, which is not the same as being right.
- It doesn't "look things up" the way a database does — unless it's been specifically connected to a live source.
- It can be confidently wrong, stating something false in exactly the same fluent tone it uses for something true. The word for this is a hallucination.
A useful way to picture it: imagine an extremely well-read, fast, eager assistant who has read almost everything — but who, rather than admit "I don't know," will occasionally invent an answer with a completely straight face. Brilliant and helpful, but never to be trusted blindly.
4. What it's good at — and where it fails
| Reach for AI when you need to… | Be careful when you need… |
|---|---|
| Draft or rewrite text | Precise facts, figures, names or dates |
| Summarise a long document | Anything very recent or time-sensitive |
| Explain a concept simply | Exact calculations you'll rely on |
| Brainstorm options and ideas | A decision you must be accountable for |
| Change the tone or format of writing | Anything involving confidential or personal data |
The pattern: AI is superb at helping you produce and shape language, and risky wherever accuracy, recency, or accountability matter — unless you check its work.
5. See it in action
"Summarise this email in two sentences: [paste a long email]"
This is exactly what AI is for. The source is right in front of you, so you can check the summary in seconds — and it just saved you five minutes.
"What was our company's revenue last year?"
Dangerous. The AI has no access to your company's private figures — so it will guess, and dress the guess up as fact. Same tool, completely different level of risk. The difference is entirely in what you asked it to do.
6. Try it yourself
5 minutes. Open your workplace AI tool. Copy a paragraph from a non-confidential document. Ask: "Summarise this in one sentence." Then check it — is the summary accurate? Did it add anything that wasn't in the original?
You've just practised the single most important habit in this programme: use it, then verify.
7. Watch-outs — start these habits today
Never paste confidential, personal, or customer data into a public AI tool.
Don't treat it as a search engine — it isn't looking facts up.
Remember it may be out of date — its knowledge has a cut-off, and it won't always tell you.
8. Responsible use
In one line
Use AI to help you think and draft — but you remain responsible for what you send, sign, or decide.
Key takeaways
- Today's workplace AI is narrow, not a mind; it predicts likely next words.
- So it produces plausible language that can be confidently wrong (hallucination).
- It's excellent for drafting, summarising, explaining and brainstorming; risky for facts, figures and decisions — unless you check.
- The golden habit: use it, then verify.