Setting up an AI helper for Mom or Grandma: a safe, simple guide
Yes, it can help: explaining a confusing letter, reminders, answering questions at 2 a.m., help writing a message. Set it up together, on their own phone, and teach three rules: never share passwords, bank details or a Social Security number; check health and money answers with a real person; and treat any urgent request for money as a scam. Keep it a helper, not a replacement for family.
What it's good for
- Explaining things
- A confusing letter from an insurer, a word in a doctor's note, how to use a new phone setting. Ask it to explain "in plain words, step by step".
- Reminders and appointments
- Many assistants can help keep track of appointments and daily routines, such as taking medication on time.
- Writing
- A thank-you note, a reply to the landlord, a birthday message to a grandchild.
- Brain games and practice
- Word games, quizzes, practising a language, talking through a crossword.
- Company, carefully
- Research suggests AI may help reduce loneliness for people who live alone. It works best alongside people, not instead of them.
Pick the right kind of app
Choose a well-known assistant from a major company, used from the phone or tablet they already have. Be cautious with apps built around an AI "friend" or romance, especially ones that push paid upgrades to keep the conversation going. And skip anything that asks for bank or account logins: banks are moving to block AI agents from customer accounts, and Chase's advice is never to share passwords.
The 30-minute setup, together
- Install it on their device and sign in with their own account.So the history and settings are theirs, and you can both find them later.
- Turn on voice if typing is hard.Most assistants let you talk and listen instead of type.
- Go through the privacy settings with them.Find where chat history and "memory" are kept, and whether conversations can be used to train the AI. Switch off what they're not comfortable with, and show them where it lives.
- Show them the thumbs-down button.It reports a wrong or strange answer to the company.
- Do three real tasks together.One letter to explain, one message to write, one question they've been wondering about. Confidence comes from using it, not from instructions.
- Put the three rules on a card by the phone.Use the one below. Add the family safe word while you're at it.
What to watch for
- Confident wrong answers
- Chatbots can sound sure and still be wrong, and can invent sources. Health, money and legal answers always get checked with a person. Why it happens: see the guide linked below.
- Leaning on it instead of people
- AARP's advice is not to use a chatbot in isolation. If it's becoming their main company, that's a reason to call more, not to take it away.
- Oversharing
- AARP says to treat a chatbot like a stranger you've just met at a bus stop: friendly, but you wouldn't give them your bank details.
- Scams that use AI
- Urgency, gift cards and "act now" mean stop. And a call in your voice asking for money may not be you: agree a safe word.
Check in, kindly
Every week or two, ask what they've used it for lately, whether anything it said seemed wrong, and whether anyone, person or bot, has asked for money or details. You're not testing them; you're keeping the door open, so they tell you when something feels off.
The card for next to the phone
Print it or copy it, fill in the blanks, and keep it by the phone or tablet.
From the record
Dated cases from the Warning Shots index and the reporting behind them.
- Lawyers filed a brief citing court cases ChatGPT had invented, then defended them. A judge fined them. Confident and wrong is the everyday risk. The case file → (Abajournal)
- A mother got a call in what sounded like her daughter's voice, crying for help. It was an AI clone. Older relatives are prime targets. The case file → (US Senate testimony)
- Banks are restricting AI agents inside customer accounts; Chase's advice is never to share passwords with them. (CBS News)
Questions people ask
Is ChatGPT (or another chatbot) safe for seniors?
It can be, with the right setup and three rules: never share passwords, bank details or ID numbers; check health and money answers with a person; and treat urgent requests for money as scams. The risks are oversharing, confident wrong answers and leaning on it instead of people.
Can an AI chatbot replace a caregiver or a visit?
No. It can help with reminders, explanations and some company, but organizations like the National Council on Aging advise presenting it as a support tool, not a replacement for relationships, care or medical advice.
Does the chatbot remember what my mom tells it?
It depends on the app and its settings. Many assistants keep chat history and can have a "memory" you can review or switch off. Go through those settings together during setup.
What if she gets attached to it?
Some attachment is normal. Watch whether it's replacing contact with people. If it is, add more human contact rather than removing the tool. Ask about it openly, without judgement.
Should I give it access to her bank or email?
No. Don't give a chatbot or AI agent bank logins. Banks are moving to block agent access to customer accounts, and the safest rule for an older parent is that money stays between them, their bank and family.
What's the single most important thing to teach?
That urgency is the alarm. Any person, voice or bot asking for money or details in a hurry means stop, hang up and call family.
Related
Sources
- AARP: 6 ways to stay safe in a chatbot world (Feb 4, 2026)
- National Council on Aging: Can AI help older adults age well? (Aug 5, 2026)
- CBS News: AI agents in customer bank accounts (Oct 23, 2026)
- Axios: Banks plan for the 'day after' an AI catastrophe (Oct 23, 2026)
Spot an error? Tell us; corrections are made and logged. Written with AI assistance from public sources; every fact checked and approved by Colin Balfe. General information, not medical, legal or financial advice.