What I Tell AI, What I Keep to Myself, and How It Watches My Back

Bonus article on using AI in everyday life.

My eldest sister called me in a panic. Our middle sister, who lives with Parkinson’s, was struggling to get the in-home care she badly needed. Since we all live thousands of miles apart, I felt helpless.

“Did you ask AI,” I asked.

“No,” she said. “I’m not getting anywhere near AI. I don’t trust it.” She sounded terrified.

I was fairly new to AI myself. But I felt it was worth exploring. So, I dove in.

I opened ChatGPT and typed: “My sister has Parkinson’s. She’s in her 70s. She needs an advocate for her care. And she needs help sorting out her insurance. She claims that she paid for in-home care throughout her career. I need someone in her area that can help. Here’s her zip code. Can you help me?”

Within moments, I had a clear, organized list: her local Area Agency on Aging (with a note to call them first), a free Medicare counseling program, the Parkinson’s Foundation helpline, and a free legal aid organization for insurance disputes. Each suggestion came with a phone number and a short explanation of how they could help.

I was amazed! And I was relieved…

I immediately started making phone calls. I left a couple of messages and reached one live person who reassured me that they could help.

That experience alone convinced me of the power of AI.

Here’s the big lesson.

Notice what I shared: her situation and her zip code.

Notice what I didn’t share: her name, her birthdate, her doctors, or her insurance numbers. ChatGPT didn’t need any of it to help me get the resources I needed for my sister.

That moment taught me something I now practice every day. AI has become one of my most trusted helpers. And like any good relationship, it works best with a little wisdom about what to share.

My eldest sister still won’t try AI. 😊

Sharing Health Questions: Give What Helps, Keep What’s Private

When I ask about health, I give AI what it needs to be truly useful:

• The condition or symptom I’m curious about
• A general age, like “in her 70s”
• The kind of help I’m seeking
• A zip code or town, only when I need local resources

And I never enter full names, birthdates, doctors’ names, Medicare or insurance numbers, or test results with anyone’s name on them.

My simple guideline: describe the role, not the identity. “My sister, in her 70s” tells AI everything it needs to know.

Receiving Health Information: A Wonderful Research Partner

This is where AI truly shines for me. It can:

• Explain a medical term in plain, everyday language.
• Help me write down good questions before an appointment.
• Point me to reputable organizations and support groups.
• Lay out options, so I walk into the doctor’s office feeling informed and calm.

What I always remember: AI is my research partner, not my doctor. I bring what I learn to my care team before changing anything. I double-check phone numbers and details on official websites, because AI can occasionally get a specific detail wrong. And I use my critical thinking. AI is not always right.

Nutritional Supplements: Take or Toss?

Over the years, I’ve spent a small fortune on nutritional supplements. Inevitably, a bottle or two ends up in the back of the cabinet or refrigerator, forgotten for months – or even years. Sound familiar?

Rather than automatically tossing them, I’ve started using AI to help me determine whether they might still be worth keeping.

I simply enter the brand, product description, and expiration date.

AI typically asks a few follow-up questions: How was it stored? Has the color, texture, or odor changed? Are there any other signs of deterioration?

Then it helps me review the manufacturer’s recommendations, available information about shelf life, and potential safety concerns.

Of course, AI can’t guarantee that an expired supplement is safe or effective. But it can help me make a more informed decision about what to keep, what to investigate further, and what to toss.

I believe this simple approach has saved me a substantial amount of money over the years!

When You’re Wondering, “What Could This Be?”

Many of us have felt it: a new ache, a strange rash, a symptom that won’t quite go away. It’s tempting to describe it to AI and ask for an answer. AI can be a wonderful place to start, as long as we know what it can and can’t do.

What to share:

• Your symptoms, in plain words: where, what it feels like, and how long it’s been going on.
• What makes it better or worse.
• A general age, like “a woman in her 70s.”
• Conditions you already live with and the types of medicines you take, described in general terms.

The more clearly you describe what you’re experiencing, the more useful the response will be. You should still leave out your name, birthdate, doctors’ names, and insurance numbers. AI doesn’t need them to help you think things through.

What to Keep to Yourself

• Lab reports, scans, or patient portal screenshots that show your name, birthdate, or record numbers. If you’d like help understanding a result, type in the values themselves or cover your personal details first.
• Photos that show your face or other identifying details, unless they truly matter to the question.

What to Expect

Possibilities, Not a Diagnosis

AI can help you connect a symptom to a possible issue. But it can’t examine you, listen to your heart, or see your full history the way your doctor can.

Answers That May Miss the Mark

AI can sometimes sound more worried than it should. Or more reassuring than it should. Don’t expect AI to give you an answer. Think of it as a thinking partner who helps you fine-tune your research or prepare for a doctor’s visit.

A Great Partner for Preparation

One of my favorite questions is: “What should I ask my doctor about this?” I walk into appointments with a short, clear list, and my visits are so much more valuable.

And one wise rule: if something feels urgent, like chest pain, trouble breathing, sudden weakness on one side, slurred speech, or the worst headache of your life, skip the AI and call 911 right away. AI is for wondering and preparing, never for emergencies.

What I Never Type Into AI

• Social Security or Medicare numbers
• Bank account or credit card numbers
• Passwords and PINs
• Photos of my driver’s license, passport, or other ID
• Other people’s private details without their blessing

None of this comes from fear. It’s the same quiet discernment I’d use chatting with a friendly stranger on a long train ride. Pleasant conversation? Absolutely. My bank PIN? Never.

I Asked AI to Look Out for Me

Here’s something many people don’t know: you can ask AI to keep an eye on what you share. I gave mine a standing instruction like this:

“If I ever start to share a password, an account number, or a Social Security number, please stop me and remind me not to.”

Most AI apps have a place in their settings for instructions like this, often called “custom instructions” or “personal preferences,” so you only need to set it up once.

One honest note: the reminder arrives after you’ve pressed send. So, I think of it as a gentle friend helping me build good habits rather than a lock on the door. Over time, the pause becomes second nature.

The Email AI Knew Was a Scam

This is an email I received recently from “other rajahpushpa123@gmail.com”:

Hello,

Your payment of $381.99 has been received through your Norton wallet.

Standard processing may take up to 12 hours. After processing is complete, the transaction will be reflected on your account statement as Norton LLC..

Subscription Details:

• Customer ID: 406222

• Subscription Plan: Norton 360 Standard • Payment Amount: $381.99

Our support team is available to help with any questions regarding your account or payment details.

If this payment was not authorized by you, please contact the billing team at +1 (864)-856-3853 within 12 hours for further assistance.

Thank you,

Member Services Team

Norton

First of all, the “From” address was very suspicious. Instead of clicking anything, I took a screenshot of the email and asked AI, “Does this look like a scam?”

ChatGPT immediately flagged it as a scam. I’m so grateful I didn’t click or reply.

The signs AI looks for are ones we can all learn to spot: a rush to act right now, a topic that effects our finances or safety, a sender address that doesn’t quite match the company, a request for payment, gift cards, or passwords, links that lead somewhere unexpected, and a greeting so generic it could be meant for anyone.

Here’s the simple routine I use now:

1. Don’t click any links or call any numbers you see in the message.
2. Copy the text or take a screenshot of the message.
3. Remove any personal details, like account numbers.
4. Paste it into AI and ask, “Does this look like a scam?”
5. If you’re still unsure, contact the company using a phone number you already know and trust. If you don’t have one, look them up online on your own. They will have contact information on their official website.

Wisdom, Gratitude, and a Helper at My Side

What moves me most is how much peace this has brought me. I can help the people I love from thousands of miles away. I can walk into appointments prepared. And I can open my inbox with confidence instead of worry.

AI didn’t replace my judgment. It gave my judgment a wonderful partner. And for that, I’m genuinely grateful.

If you’ve found this article helpful and would like to learn more, grab a copy of my free guide 10 Fun (and Safe!) Ways Seniors Are Already Using AI.

My partner, Tom Lemke, and I are teaching a series called AI Made Easy for Adults 60+. We are both in our 70s and use AI to share wisdom with the world. The next series of four weekly hands-on classes starts Thursday, October 29. You can learn more at AI Made Easy for Adults 60+.

Let’s Have a Conversation:

What are your main fears regarding AI? If you have used AI apps before, what was it for – conversation, research or something else? Please share with the community!