AI for beginners over 50: a calm first week
If you're over 50 and have never used ChatGPT or any other AI assistant, you're not behind. You're in the majority. In a February 2026 Pew Research Center survey of US adults, 58% of people aged 50 to 64, and 77% of those 65 and over, said they never use AI chatbots.
The same survey asked non-users why. The most common reasons, apart from simply not being interested, were worries about how their personal information would be used (79%), not trusting the answers to be accurate (76%) and not knowing how to use them (55%). Those are sensible worries, and this plan deals with each of them.
It takes one week, about 20 minutes a day. You can stop after any day. You don't need to buy anything.
Before you start: what you need
- A smartphone, tablet or computer with internet.
- An email address you can get into.
- A notebook and pen. Writing things down helps, and you'll want a record of what you set up.
- Optional: a family member or friend who already uses one of these tools. Learning together is easier, but you can do this alone.
Day 1: Choose one assistant and set it up safely
There are several well-known AI assistants, including ChatGPT, Claude, Grok and Gemini. All have free versions, and for a first week it doesn't matter which you pick. If someone in your family already uses one, pick that one so they can help.
- Go to the official address only. Type it carefully: chatgpt.com, claude.ai, grok.com or gemini.google.com. On a phone, install the app from the official app store and check the maker's name. Look-alike apps with similar names exist.
- Create an account with your email and a password you don't use anywhere else.
- Open the settings and look for two things: chat history, and whether your conversations can be used to improve the service. You can usually switch the second one off. If the settings look confusing, ask a trusted person to look with you.
- Write down which assistant you chose, how you open it and which settings you changed.
That's enough for Day 1.
Day 2: Your first conversation
Type a question the way you'd ask a patient, well-read friend. There are no special words. Try one of these:
Explain what a website "cookie" is, in simple words, and whether I should accept them.
Suggest three easy dinners for one person using eggs, potatoes and spinach.
Read the answer. Then reply to it, just as you would in a conversation: "Make that shorter", "Explain the second part again" or "I don't have spinach, what else could I use?" The assistant remembers what you said earlier in the same chat, so you can keep refining.
The thing to notice: you can ask follow-up questions as often as you like. Nobody is judging, and nothing you type in this kind of chat is sent to anyone else as a message.
Day 3: Ask it to explain something you've been putting off
Most people have a letter, a bill or a new gadget they don't quite understand. AI assistants are good at explaining things in plain words. Try:
My new phone keeps asking me to turn on "two-factor authentication". What is it, is it worth doing, and what will I need?
Explain the difference between a fixed-rate and a variable-rate savings account, in plain English, with a simple example.
If you want help with a real letter, type out the general question, not the private details. Leave out names, account numbers and addresses. "My energy supplier says my direct debit is going up because of an 'annual review'. What does that usually mean, and what questions should I ask them?" gets you a useful answer without sharing anything personal.
Day 4: The "never share" list
This is the most important day. When you type into an AI assistant, your words go to the company that runs it and may be stored. Treat it like talking to a helpful stranger: friendly and useful, but not someone you hand your private papers to.
Never type in:
- Passwords, PINs or the codes banks text to you
- Bank or card numbers
- National ID, passport, tax or social security numbers
- Your full address together with your name
- Medical details that identify you or someone else
- Other people's private information
If you need help with something personal, describe it in general terms, or swap real details for placeholders like [my bank] or [my neighbour]. You'll get the same help.
Write this list in your notebook. It's the one page worth keeping by the computer.
Day 5: Check before you trust
AI assistants can be wrong, and they sound just as confident when they're wrong as when they're right. They can invent facts, get numbers mixed up and even make up sources. The fix is a simple habit:
- For anything that matters (health, money, legal questions, travel times, phone numbers), check the answer against an official source: your doctor or pharmacist, your bank's own website, a government site.
- Ask the assistant where the information comes from, then look at that source yourself.
- Treat it as a starting point, never the final word. It's excellent for explaining and drafting. It isn't a doctor, lawyer or financial adviser.
Try it now: ask a factual question you already know the answer to, such as the opening times of your local library or the dose instructions on a medicine you take, and compare. You'll quickly get a feel for when it's reliable and when it isn't. Our guide on how to check an AI answer has more on this.
Day 6: Let it help you write
Writing is where many people find AI most useful. Try one of these:
Help me write a short, warm thank-you note to my neighbour, who watered my plants while I was away for two weeks.
I need to email a shop because the kettle I bought stopped working after a month. Help me write a polite, firm message asking for a replacement. Keep it under 120 words.
Read the draft, change anything that doesn't sound like you, and remember: you are the one who sends it. The assistant only drafts. If you'd like it to sound more like you, tell it: "Make it less formal" or "I'd never say 'kindly', use plainer words."
Day 7: Know how scammers use AI, and agree a family code word
The same technology that writes a thank-you note can help criminals. The US Federal Trade Commission warns that scammers can clone a family member's voice from a short clip posted online, then call pretending to be in trouble and needing money urgently.
Three rules protect you:
- Don't trust the voice alone, even if it sounds exactly right. Hang up and call the person back on a number you already have.
- Be suspicious of urgency and secrecy. "Don't tell Mum and Dad" and "I need it in the next hour" are warning signs. So are requests for gift cards, wire transfers, cryptocurrency or payment apps.
- Agree a family code word. Choose a word or phrase with your family that a stranger couldn't guess. If someone calls in an emergency and can't give it, treat the call as a scam. The FCC's guide to grandparent scams recommends the same idea.
Use today's 20 minutes to ring or message your family and agree the word.
After the first week
You've chosen a tool, set it up carefully, had real conversations, learned what never to share, practised checking answers, and protected yourself against the most common AI scam. That's more than most people do.
From here, use it when it's genuinely helpful: planning a trip, understanding a confusing form, learning a new hobby, adapting a recipe. If a week goes by and you don't use it, that's fine too. It will still be there.
Want it all in large print?
AI for Beginners Over 50 is our printable starter guide: 11 pages in 13pt large print, in plain words, with space for notes. It covers what an AI assistant is and isn't, step-by-step safe setup, five safe first things to try, what never to type and what to do instead, checking answers and avoiding AI scams (including the family code word), a table of everyday uses with a two-week plan, and a quick reference card. It also makes a thoughtful gift for a parent or grandparent who's curious but cautious.
Using AI at work as well? The AI Safety & Privacy Checklist for Work turns the never-share list and checking habit into a one-page policy for a team.
Not ready to buy? Start with the free guide, Your first week with AI: ten prompts and a privacy checklist, plus a five-day email course.