Get real help from AI — and you stay in charge.

Welcome

This guide is for anyone who is new to AI, curious about it, and a little worried. You don’t need to be good with computers, and you don’t need to pay for anything. The free versions of these AI helpers are plenty to start with.

You can’t break anything by asking a question. Ask anything, as many times as you like.

In about 15 minutes, you’ll know how to set up an AI helper safely, keep your information private, catch it when it’s wrong, and spot the scams that use it.

You’re not alone. More than half of adults over 50 have used AI, counting voice helpers like Siri and Alexa. Half aren’t confident they could tell when AI gets something wrong — that’s normal, and it’s why this guide exists. And 9 out of 10 want to know when something was made by AI, just like you.

How to use it. Each section is short and stands on its own. When you see words set off in a box like this one, those are the exact words to type:

Want to try something right away? Jump to Section 6. The cheat sheet at the end prints front and back. Keep it by your computer.

1. What AI is, and what it isn’t

An AI helper is a program you talk to by typing or speaking. You ask a question in everyday words, and it answers in everyday words. ChatGPT, Google Gemini, Microsoft Copilot, Claude and Muse are the best-known ones.

Things you might not know it can do:

• Look at a photo and explain it. Snap a picture of a menu, a label, a plant or a confusing sign, and ask what it means. Section 6 shows you how, step by step.
• Talk with you out loud, like a phone call, so you don’t have to type.
• Turn handwriting into typed text, like an old recipe card or a family letter. Just send a photo of it.
• Read something long for you, like a contract or a manual, and tell you what matters.
• Help you translate a conversation, back and forth.
• Help you tell your story, with questions to ask the grandkids or help writing down family memories.
• Make pictures from a few words. ChatGPT, Gemini, Copilot and Muse can. Claude can’t.

Most AI helpers can do most of these for free, though free versions limit how many photos and pictures you get each day.

What it isn’t:

• It isn’t a person. It isn’t your friend, even when it sounds like one.
• It isn’t always right. It can sound completely sure and still be wrong. Section 4 shows you how to check.
• It isn’t a doctor, lawyer or financial advisor. Use it to understand things and prepare questions, then ask a real professional.
• It isn’t a safe place for secrets. Section 3 covers what never to type in.

Think of it as a very well-read helper who sometimes makes things up. Useful, as long as you stay in charge.

You set the limits.

• It advises. You decide.
• It never sends, buys, or books anything as you without showing you first and getting your okay.
• Start with small jobs. Let it earn bigger ones, same as a person.
• If it agrees with everything you say and praises everything you do, be careful. A good helper tells you the truth.

You can talk instead of type. Most AI helpers have a microphone button. Tap it and speak, and many can read their answers out loud, too.

Stuck? Ask the AI itself. It can explain its own buttons and settings.

2. Set up your AI helper safely

Pick one AI helper and learn it well. The easiest choice is the one that comes with what you already use: Gemini on an Android phone or with Gmail, Copilot on a Windows computer. ChatGPT, Claude and Muse work on anything.

Step 1: Get it from the real source. Download the app only from your phone’s official app store, and check that the maker is the real company: OpenAI, Google, Microsoft, Anthropic or Meta. On a computer, type the address yourself: chatgpt.com, gemini.google.com, copilot.com, claude.ai or muse.ai. Fake look-alike apps and websites exist.

Step 2: Make your account. Sign up with your email address, or with an account you already have, like Google or Microsoft. Choose a strong password you don’t use anywhere else. The free versions don’t ask for a credit card. If it asks for payment, you’ve picked a paid plan, so stop and go back. One exception: Muse. Muse asks for a card when you sign up, even for the free version. You aren’t charged unless you upgrade.

One fake to watch for: an app named Muse: AI Video, Photo Generator — made by Mars Studio — is not Meta’s Muse. The real one is Muse from Meta; its maker line says Meta Platforms.

If you ever see “You’ve reached today’s free limit,” nothing is broken and you did nothing wrong. The free version just ran out for today — it resets tomorrow. And you never have to pay to keep anything you already made.

Step 3: Stop the company from learning from your chats. By default, these companies may use what you type to teach and improve their AI. This is called “training.” Here’s how to turn it off:

AI helper	Where to go	What to turn off
ChatGPT	Settings, then Data controls	Improve the model for everyone
Gemini	Settings & help, then Activity	Keep Activity
Copilot	Your profile, then Privacy	Model training on text, and Model training on voice
Claude	Settings, then Privacy	Help improve our AI models
Muse	Settings, then Data Controls	Turn off using your chats to help develop Meta’s AI

Product details checked October 2026 — apps change their menus, free limits and abilities often.

These menus are for a computer, and a phone may differ. If you can’t find a setting, ask the AI: “Where is the setting to stop using my chats for training?” Turning it off doesn’t erase your chats, so Section 3 still matters.

In Gemini, turning off Keep Activity also means it won’t keep a list of your past chats. That’s normal.

Step 4: Try one small, harmless question.

Read the answer. Ask a follow-up. That’s all there is to it.

Last stop in setup: meet the chat box

You just used two parts of it — the typing box and the send arrow. Every helper’s screen has the same few pieces, even though they move around a little from app to app. Learn them once and every AI app will feel familiar.

Hello! Ask me anything.

Copy — under every answer

+ Type your words here The microphone

The + button The typing box The send arrow

The typing box — the long bar, usually at the bottom. This is where your words go. Tap it and type, the way you’d type a text message.

The + button — in or beside the typing box, usually at the left. This is the add button. It’s how a picture or a file gets into the chat. (Some apps show a paperclip instead of a +. Same job.)

The microphone — tap it and talk instead of typing. Tap it again when you’re done.

The send arrow — the round button at the right end of the box. It sends your message. On a computer, the Enter key does the same thing.

How to add a picture, exactly:

1. Tap the +.
2. Tap Photos (it may say Photo Library). Your pictures open.
3. Tap the picture you want. It appears just above the typing box.
4. Type what you want to know — like: What does this say?
5. Tap the send arrow.

On a computer it’s the same idea with clicks: click +, click Upload or Photos, pick the picture, type your question, press Enter.

A document goes in the same door: tap +, then Files, and pick the letter or document instead of a picture. Added the wrong one? Before you send, tap the little × on it to take it back out.

Under every answer there are small buttons. The one you’ll use most is copy — it looks like two little pages, one behind the other. Tap it and the whole answer is yours to paste into Notes, a message, or an email. That’s how a letter your helper wrote gets where it was going.

The door: starting a new chat. Each chat is its own conversation, like a separate phone call. For a fresh one, there’s a New chat button — often a pencil picture at the top, or “New chat” in the list at the side.

Start a new chat for a new subject: one chat for the electric company letter, another for recipes. It keeps things from tangling.

And if a long chat gets muddled and the answers start missing what you said, don’t fight it. Start a new chat and repeat the one fact that matters: This letter is from the electric company, and I want to know what it wants from me. Your old chats stay in the list at the side, so you can go back to any of them.

One thing to carry into the next section: whatever you add with that + button goes to the AI company, same as anything you type. So the question is never just how to share something — it’s what’s safe to share. That’s Section 3.

3. Keep your personal information safe

If you wouldn’t write it on a postcard, don’t type it into an AI.

Your chats are stored on the company’s computers, and you don’t control who sees them.

Never type in:

• Your Social Security or Medicare number
• Bank account, credit card or PIN numbers
• Passwords of any kind
• Your full medical records
• Private details about other people

You can still get help with personal paperwork. Cover the private parts first. Type XXXX in place of account numbers, and leave out your name and address. The AI doesn’t need them to help.

What you can share. You don’t have to hide your whole life. Sharing the story of your life is fine: your first name, your town, your hobbies, a bad knee in plain words, your grandkids’ first names for a birthday card. Details like these help the helper help you. Just never share the keys — numbers, passwords, anything that could unlock an account or let someone pretend to be you.

Here’s the test. Ask: could a stranger with this detail take my money, or pretend to be me? If yes, never type it. If it only tells the helper about your life, it’s fine.

One more control. Some helpers remember what you tell them, from one chat to the next. Ask:

Read what it says. Anything you don’t like, tell it: “Forget that.”

Two more habits:

• Delete chats you don’t need. Every AI helper has a way to delete a conversation, usually a trash can or three dots beside it.
• Be careful with photos. A picture of a document can show more than you meant, like an address, a signature or a number in the corner. Crop or cover those parts before you share it.

If you ever type something private by mistake, delete that chat right away. If it was a bank or card number, call your bank.

4. How to tell when AI is wrong

AI sometimes makes things up and says them with total confidence. In a 2025 national poll, half of adults over 50 weren’t confident they could tell when AI gets something wrong. These checks take a minute and catch most mistakes.

Check anything about health, money or the law with a real person.

Be extra careful when the answer:

• Gives exact numbers, dates, names or quotes
• Is about your health, money or a legal matter
• Is about something recent, like prices, rules or news
• Agrees with whatever you already believed

Four quick checks:

1. Ask where it came from. Then open the link and see if the page really says that.

   Where did you get that? Give me a link I can check.

2. Ask how sure it is.

   How sure are you? What part of this could be wrong?

3. Ask a second source. Put the same question to a different AI helper, or look it up on an official website. If they disagree, check with a person you trust.
4. Check anything important with a real person. That means your doctor, your pharmacist, your bank, or the agency itself.

Tip: tell it not to guess. Say this at the start of a chat, and it will make fewer things up:

For health questions, use AI to understand words and prepare questions for your doctor, not to decide on treatment.

And when you want a real person: your public library gives free help with phones, computers, and AI — real people, no charge, no judgment. Bring this guide.

5. How to spot AI scams and fakes

Scammers now use AI to copy a loved one’s voice and to make fake videos that look real. The good news: a few simple habits beat even the best fake.

When someone asks for money, hang up and call back on a number you know.

The warning signs:

• A rush. “Act now,” “don’t hang up,” “before it’s too late.”
• A request for money by wire, gift cards or cryptocurrency. A real loved one won’t pressure you to pay this way.
• A request for secrecy. “Don’t tell Mom.”
• Something slightly off in a video, like strange hands, odd eyes, or lips that don’t match the words.

What to do:

1. Set up a family secret word today. Pick a word only your family knows. If someone calls in trouble, ask for it. The FBI recommends this.
2. Hang up and call back on a number you already know. Do this even if the voice sounds exactly right.
3. Never send money, gift cards or crypto to someone you’ve only met online or by phone.
4. Keep your social media private, so scammers have fewer voice clips and photos of you to copy.

When in doubt, hang up and call back. If you’ve been scammed, report it at reportfraud.ftc.gov.

6. Copy, paste, done

Copy any of these into your AI helper. Change the words in [brackets] to fit your life. Remember Section 3: cover private numbers with XXXX.

Start every chat with this line. It keeps answers short and plain, and cuts down on guessing:

The recipe for asking for anything

1. Say what you want. “Write an email…”
2. Give the details. Who, what, when.
3. Say how you want it. Short, friendly, a list, plain words.
4. Fix it together. “Make it shorter.” “Make it friendlier.” “Try again.”

Writing

An email

A thank-you note

A letter to a company

A birthday message

Lists

Groceries

A to-do list

Packing for a trip

Documents

A summary of something long

A sheet for a doctor’s visit

A simple budget

Dates and reminders

Some AI helpers can add things to your calendar and some can’t. These commands work either way: the AI shows you first and waits for your OK, or tells you how to do it yourself.

An appointment

A birthday every year

Everyday help

A confusing letter or bill

Learning something new

Making a choice

A scam check

Screenshots

Show it your screen. Your helper can see, not just read. So when you’re stuck, don’t describe the screen — show it.

1. Take a screenshot. iPhone: Side button + Volume Up, together. Android: Power + Volume Down. On a computer, ask your helper how — it depends on the computer. The picture saves to your Photos or Pictures, newest first.
2. In your helper’s app, tap the + or the paperclip, and pick the picture.
3. Ask:

   What does this mean, and what do I tap next? One step at a time.

4. Do the step. New screen? Screenshot again and keep going.

Some helpers can even look at your screen live while you talk. Ask yours: “Can you see my screen?”

Turn a picture into a table or spreadsheet

1. Photograph the page straight-on — whole page, good light. A bill, a receipt, a handwritten list.
2. Send it and ask:

   Read everything in this picture and make it a neat table: item, amount, date.

3. Check the total first, then a number or two against the picture. A helper can misread a digit.
4. Shape it with plain orders: “Sort it by amount.” “Make it a spreadsheet I can download.”

Before you send any picture: crop it, or cover account numbers, card numbers, and passwords. Same postcard rule as Section 3.

Have some fun

A funny picture

A meme

Pictures work in ChatGPT, Gemini, Copilot and Muse, with a daily limit on the free versions. Claude can’t make pictures. Muse can make short videos; with the others, video usually takes a paid plan.

Label what AI made. Before you share a picture, add a note like “AI image.” Never make fake pictures of real people.

When an answer isn’t right, just say:

• “Make it shorter.”
• “Make it friendlier.”
• “Use simpler words.”
• “Put that in a numbered list.”
• “Try again, but [what to change].”

Cheat sheet: keep this by your computer

Front: Copy these

Start here

• “I’m new to AI. Use plain words, keep it short, and ask me if you need more information.”
• “What can you do for me? Give me 10 examples.”
• “Walk me through it one step at a time.”

Get things done

• “Write a short, friendly email to [who] about [what].”
• “Make a list for [groceries / my trip / today].”
• “Help me put [this] on my calendar. Show me first.”
• “What does this letter mean? Do I need to do anything?”
• “Does this look like a scam?”
• “Make a funny picture of [anything].”

Fix the answer

• “Make it shorter and simpler.”
• “How sure are you? Where did you get that?”

Back: Stay safe

Three rules

1. If you wouldn’t write it on a postcard, don’t type it into an AI.
2. Check health, money and legal answers with a real person.
3. Someone asks for money? Hang up and call back on a number you know.

Never type in: Social Security or Medicare numbers, bank or card numbers, passwords.

Sources

• Data controls FAQ, OpenAI
• Gemini Apps activity help, Google
• Microsoft Copilot privacy controls, Microsoft
• How do I change my model improvement privacy settings, Anthropic
• How to manage your Muse data, Meta
• Family emergency scams, Federal Trade Commission
• Criminals use generative AI to facilitate financial fraud, FBI public service announcement, December 3, 2024
• How older adults use and think about AI, University of Michigan National Poll on Healthy Aging, 2025
• More Americans are using AI for health questions, AARP, June 2026

App menus change often. If a setting has moved, search the app’s help pages for the setting’s name.

The Grocery List

Ends with: a grocery list for the week’s dinners, grouped by parts of the store.

The Letter Answer

Ends with: a short reply you have heard read back to you, ready to send yourself.

Before you photograph a letter: cover the account number with a scrap of paper. If you wouldn’t put it on a postcard, don’t send it to a helper.

The Spreadsheet

Ends with: a tidy spreadsheet of your list, sorted, with a total — saved as a file when the helper can make one.

Free versions limit how many files they’ll make in a day. If your helper says you’ve hit a limit, ask it to show the list right here in the chat instead.

The Doctor Visit

Ends with: one page of questions to take to your doctor. This set only helps you get ready — it never guesses what’s wrong. A real doctor does that.

The Thank-You Note

Ends with: a short, warm note in your own voice, heard back before it goes anywhere.

The Day Plan

Ends with: your day in a short, sensible order you can check off, one thing at a time.