What AI actually is (and isn't)
You've heard the words everywhere — ChatGPT, AI, "artificial intelligence" — usually wrapped in either breathless excitement or vague dread. Let's strip all that away and describe, in plain English, what one of these things actually is. It's far less mysterious than the headlines suggest.
What it is
An AI assistant like ChatGPT or Claude is, at heart, a remarkably good writing and conversation tool. You type a question or a request in ordinary English, and it types back a sensible, well-written reply. Think of it as an extraordinarily well-read assistant who has read a vast amount of text and is very good at putting things into words — available any time, with infinite patience, and never bored by your questions.
You don't need any special language or technical skill. If you can send a text message or write an email, you already have everything required to use it.
What it isn't
It is a tool, not magic, and not a person. It doesn't "know" things the way you do, and it has no opinions or feelings of its own, however human it may sound. Most importantly: it is not always right.
The right mental model
Picture a clever, patient assistant who's read almost everything but occasionally misremembers and never admits it. Wonderful for getting started, drafting and explaining — but you're still the one in charge, and you check the important details. Hold that picture and you'll use it well.
Your first conversation
Enough theory. The best way to understand this is to try it, and it's free to do so. Let's get you talking to one.
Getting to a free one
- Open your web browser and type the address yourself: chatgpt.com or claude.ai. Both have a free version that's perfectly good for everything here.
- Create a free account with an email address and password, or sign in with a Google account if you have one.
- You'll see a simple box at the bottom of the screen that says something like "Message" or "Ask anything." That's where you type. Press Enter to send.
How to ask well
The single skill worth learning is how to ask. The trick is simple: give it context and be specific. A vague request gets a vague answer; a clear one gets a genuinely useful reply. Tell it who you are, what you want, and how you'd like it.
Instead of "write a letter," try: "Help me write a polite but firm letter to my energy supplier. I'm 68, I've been overcharged for three months, and I want a refund and an apology. Keep it short." See the difference? The second version gives it everything it needs.
Three worked examples to try right now
It's a conversation, not a one-shot
You don't have to get it perfect first time. If the reply isn't quite right, just say so: "shorter," "less formal," "explain that bit more simply." It remembers what you've been discussing and adjusts. Going back and forth is exactly how it's meant to work.
Useful everyday tasks
Once you're comfortable chatting to it, the real value appears: small, genuinely useful jobs it handles in seconds. Here are some recipes worth keeping. Copy the wording, swap in your own details.
Recipes worth keeping
That last one is where these tools quietly shine. They never tire, never sigh, and never make you feel slow. You can ask the same question three different ways until it clicks. For picking up a new hobby, learning to use your phone, or finally understanding something you were always too embarrassed to ask about, a patient companion that's available at midnight is a genuine gift.
Staying safe — read this twice
Three rules protect you completely. First: never share passwords, bank details, PINs or full personal information with an AI — it doesn't need them, and you should be wary of anything that asks. Second: AI can be confidently wrong, so always verify anything important — a medical, legal, financial or factual matter — with a real, qualified source before acting on it. Third: be alert to scam calls using AI-generated voices that imitate a family member in trouble and ask for money urgently. If you ever get such a call, hang up and ring the real person back on their usual number. Agree a family "safe word" now, just in case.
Going deeper
Once the basics feel natural, you can explore further — at whatever pace suits you. There's no race here, and no shame in sticking with just the bits you find useful.
First conversation — you're here
Week 1You've signed up and asked your first few questions.
Everyday helper
Month 1You reach for it naturally for letters, summaries and explanations.
Confident user
Month 2–3You give good context, push back when it's wrong, and trust your own judgement.
Comfortable and safe
OngoingIt's a useful tool in your kit — used wisely, never blindly.
Other tools worth knowing
Beyond ChatGPT and Claude, there's Google's Gemini (built into many Android phones) and Microsoft's Copilot (in Windows and Office). They all work in much the same way, so the skills you've learned here transfer directly. You don't need to learn them all — pick one you like and stay with it.
Staying safe and private
A few sensible habits: don't type anything into an AI you wouldn't be comfortable sharing more widely; remember that conversations may be stored; and keep that family safe word in mind for any urgent call asking for money. Caution here costs nothing and protects a great deal.
Keeping up without being overwhelmed
The technology changes fast, and you do not need to chase it. Learn one tool well, ignore the hype, and add new tricks only when you have a real use for them. Being a calm, careful, occasional user is perfectly fine — better, in fact, than being a frantic one.
- Age UK digital guides — Friendly, jargon-free help getting started with technology, written for older adults. ageuk.org.uk.
- BBC and Which? technology pages — Trustworthy, plain-English explainers on new tools and how to stay safe with them.
- Age UK
- Tech for Seniors
- ageuk.org.uk
- chatgpt.com
- claude.ai
- ChatGPT (free assistant)
- Claude (free assistant)
- Google Gemini (often built into Android)
