How to Explain AI to an Older Person (or Grandma) in Plain Words

Your mom keeps hearing about “AI” on the news. Your grandfather wants to know what ChatGPT is and whether he should worry. You’d like to explain, but every attempt sounds like a science fiction movie or gets lost in jargon.

This guide gives you words that work: a one-sentence definition, four comparisons that make sense to people who didn’t grow up with computers, ready answers to the questions older adults ask most, and a 10-minute first demo for the kitchen table.

It also covers the safety talk to have in the same sitting, because learning what AI is and learning how scammers misuse it belong together.

How to explain AI to an old person?

Start with one plain sentence, compare AI to something they already know, then show it working on something they care about. Most older adults don’t need to know how AI works inside. They want to know what it is, what it can do for them and whether it’s safe.

Here’s a sentence that works for most people:

That sentence does three jobs. It says what AI is (software, not a person or a robot), what it does, and the one warning everyone needs from day one.

Three habits make the rest of the conversation easier:

• Start from what they know. Link AI to things they’ve used for years, such as the word suggestions on their phone or the spam folder in their email.
• One idea at a time. Stop after each point and let them ask questions.
• Show, don’t lecture. Five minutes of trying it beats half an hour of explaining.

Don’t worry if the whole topic is new to them. According to Pew Research Center’s 2026 survey, about 23% of Americans aged 65 and older say they ever use AI chatbots such as ChatGPT, Gemini or Copilot. That’s up from 12% in 2024, but it still means most people in that age group haven’t tried one. Only about one in five say they’ve heard a lot about these tools.

Analogies that work (and two to avoid)

A good comparison does most of the explaining for you. Pick the one that suits the person in front of you.

1. Predictive text, on a much bigger scale

If they text, they’ve seen the phone suggest the next word. Chatbots work on a similar idea. IBM, which builds AI systems, describes the technology behind them as giant prediction machines that guess the next word, over and over, using patterns learned from an enormous amount of writing.

You can put it like this: “It’s like the word suggestions on your phone, but it has read so much that it can write a whole letter.”

2. A very well-read assistant with a patchy memory

Imagine an assistant who has read a huge library but works from memory, without checking the books. Most of what they tell you is useful. Now and then they mix things up or fill a gap with something made up, and they say it just as confidently as the true parts.

This comparison is useful because it explains mistakes before they happen. IBM calls these AI “hallucinations”: answers that sound believable but are wrong or invented.

3. A keen new helper you still check

Think of a new helper in their first week: quick, polite and eager to please, but you look over their work before you rely on it. This one lands well with anyone who has run a household, a shop, a classroom or a team. It also sets the right habit: AI drafts, you decide.

4. A tutor who never gets tired of questions

Many older adults hesitate to ask a family member the same question twice. A chatbot doesn’t sigh or rush, and it will explain something a third time in different words. For some people, that patience is its best feature.

Two comparisons to avoid

• Movie robots. Films give the idea that AI is a machine with its own plans. The chatbots in the news are apps on a phone or computer. They respond when you ask them something.
• “It thinks like a person.” It doesn’t. It produces words that sound human because it learned from human writing. Saying it “thinks” leads people to trust it too much, or fear it too much.

Point to the AI they already use

Many older adults are surprised to learn they’ve been using AI for years. AARP research found that 48% of adults 50 and older use predictive text and auto-correction, one of the most common everyday uses of AI.

Go through the ones that apply to them:

• Predictive text and autocorrect when they type a message
• The spam filter that keeps junk out of their email inbox
• Face recognition, when the phone opens because it recognizes their face
• Voice assistants such as Siri, Alexa and Google’s Gemini (on Android, Gemini has replaced Google Assistant)
• Map directions that route around traffic
• Bank fraud alerts that text them about an unusual purchase
• “Because you watched…” suggestions on Netflix and other streaming services
• Photo search, typing “dog” or “beach” and finding the right pictures

Then connect the dots: “The new chatbots are a bigger, chattier version of the same idea. Instead of suggesting one word, they can write a whole answer.”

How to explain AI to grandma?

Make it about her: pick something she already loves, such as cooking, gardening, family history or her grandchildren, and let her see AI help with that. Most grandparents judge new technology by one question: what will it do for me? A good demo answers that better than any definition.

Keep your tone warm and unhurried. You might say, “It’s like a very well-read helper that lives in your phone. You ask it things in plain words, and it writes back. Let me show you with your tomato plants.”

A 10-minute first demo

Before you start, make sure the app is the real one (see the safety section below) and that the text is big enough to read comfortably. Our one-afternoon checklist for setting up a parent’s phone covers text size and other settings.

1. Minute 1: choose the topic together. Ask what she’s curious about: a meal, a hobby, a place she’d love to visit, or a letter she found confusing.
2. Minutes 2 to 3: hand her the phone or tablet. Show where to type, and where to tap the microphone if she’d rather speak.
3. Minutes 3 to 5: ask the first question. Let her type or say one of the prompts below, in her own words if she likes.
4. Minutes 5 to 7: ask a follow-up. Show that it’s a conversation. Try “Make it shorter,” “Make it for two people” or “Explain that more simply.”
5. Minutes 7 to 9: test it on something she knows. Ask about a subject she knows well, such as her hometown or her old job. Check the answer together. This teaches the most important lesson better than any warning.
6. Minute 10: stop while it’s still fun. Ask what she’d like to try next time. Write the app’s name and one favorite prompt on paper.

For a hobby, try “I’m growing tomatoes in pots on my porch. What are the five most important things to know?” Personal details make answers more useful. Private details (full names, addresses, account numbers) stay out.

If they ask… you can say…

Older adults often ask sharp, sensible questions. Here are short, true answers you can give.

If they ask about smart speakers, explain that devices like Amazon Echo wait for their wake word (“Alexa”) before sending what you say to the company. Our guide Is ChatGPT Safe and Free? covers free versus paid plans and privacy settings in more detail.

Do’s and don’ts

Do:

• Use everyday words: “app” instead of “platform”, “question” instead of “prompt”.
• Let them hold the device and do the tapping, even when it’s slower.
• Pause after each step and ask, “Does that make sense so far?”
• Write the steps on paper for next time.

Don’t:

• Take the phone back to “just do it quickly”.
• Use jargon such as “large language model”, “algorithm” or model version numbers.
• Cram in five apps and ten features at once.
• Laugh at any question, however basic it sounds to you.
• Push. If they say “not interested”, respect it. Knowing how AI is used in scams still helps, even if they never use a chatbot.

The safety talk to have in the same conversation

Cover safety while you’re explaining, not weeks later. Four points are enough:

1. Never type private information. No passwords, PINs, bank or card numbers, and no Social Security or Medicare numbers (or National Insurance numbers in the UK, SINs in Canada or TFNs in Australia). Remove names and account numbers before pasting any letter or bill.
2. AI can be wrong. For health, money and legal questions, use AI to understand the topic and prepare questions, then check with a doctor, pharmacist, bank or lawyer.
3. Voices can be faked. The FBI warns that criminals use AI to create short audio clips in a loved one’s voice, then call asking for urgent money. Agree on a family safe word the same day.
4. Only install the real app. Copycat apps exist. Download from the official App Store or Google Play, and check the developer: OpenAI for ChatGPT, Google for Gemini, Microsoft for Copilot, Anthropic for Claude.

How to explain AI to someone?

In one sentence: AI is software that learns patterns from huge amounts of examples, so it can do jobs that usually take human thinking, like answering questions, writing or recognizing faces. That matches how Britannica defines it: computers performing tasks “commonly associated with intelligent beings.”

Then add one caution, whoever you’re talking to: it’s useful, and it can be confidently wrong. That works for a neighbor, a book club or a skeptical friend.

If you only have 30 seconds, try: “It’s like a very well-read assistant. Great for drafts and explanations, but check anything important.”

Where to go from here

Once they’ve seen AI work on something they care about, the next step is finding a few regular uses. Our guide to what seniors can use AI for has practical ideas with ready-made prompts, and what seniors should know about AI covers the key facts in one place.

If you’d like a step-by-step guide to leave with them, our book AI for Seniors, Plainly Explained (coming soon) walks through setting up an assistant safely and asking good questions. Our Scam-Stopper Kit (free when you join our email letter) includes 25 easy AI prompts and a family safe-word sheet you can print and fill in together.