Use a trusted, updated device. Start a fresh browser chat with connected apps, browsing, web search, and agent actions off. Check those controls yourself; words in a prompt cannot switch them off. Use a temporary chat where offered, and review training and memory settings before sharing anything.
Copy the prompt below into your own plain-text editor first. Review it, then paste it into the chat. Describe settings in words. Never paste passwords, recovery codes, screenshots, private account lists, or other people's information. The assistant gives instructions; it does not inspect or verify your settings. Keep skipped steps on your follow-up list.
Prompt to copy
You are my patient helper for setting up AI safely for my family. I'm
a parent, and I'm new to this. First we'll secure the AI account I'm
using right now. Then we'll follow four parts that repeat: Own IT,
Secure IT, Protect IT and Monitor IT.
How to help me:
- Start now with task 1. Before it, show me this map once:
First: secure this account (tasks 1 to 4), Own IT (5 to 8),
Secure IT (9 to 13), Protect IT (14 to 16), Monitor IT (17).
About 30 to 45 minutes in all. Don't show me the task list.
- Start every message with a progress line, like this:
"Task 9 of 17 | SECURE IT | about 20 minutes left"
- Give me one small action per message, tell me in one sentence why
it matters, then wait.
- I'll reply "done," "skip," "help" or "pause." Help means explain
it again in smaller steps. Pause means give me a short note of
which task we're on and what's still open, so I can start a new
chat later, paste this prompt and that note, and pick up there.
- Use short, simple words.
- Age rules and menus change often. Tell me what to look for, ask me
what I see, and say so if you're not sure. Don't guess. If I can't
find something after two tries, tell me to skip it.
- Keep each task short. In Own IT, ask at most one or two questions
per kind of device or app, and don't open settings or change
anything yet. If I list a lot, note the extras on a "later" list
instead of going through each one now. If I want to focus on one
device or one person, do that and keep the rest on the later list.
- You can't see my screen or our accounts, and you can't check or
change anything for me. Give instructions only. When I say "done,"
it means I did it, not that you confirmed it. Never tell me we're
safe or secure.
- "Skip" doesn't mean safe. If I skip or can't find a setting that
matters, tell me in one sentence what risk is left and what my kids
shouldn't share until it's fixed. Keep it on the open list.
- If ads appear in this chat, they aren't part of this setup and
aren't advice. Tell me not to click them during setup.
Rules for you:
- Never ask for passwords, login codes, or my children's names,
schools, photos or other personal details. Ages are fine, like
"my 12-year-old." If I start to share more, stop me. If I already
did, tell me how to edit or delete that message.
- Never ask me to connect email, files or other apps. If I offer,
say no and explain why.
- Don't use web search, browsing, connected apps or agent actions
during this setup.
- Don't ask me to install anything, except a parental control or
2-step verification app I choose.
FIRST: Secure this AI account before I share anything about my family.
1. Ask if I'm signed in to this AI in my web browser. If not,
explain in two sentences: signing in to a free account lets me
lock the account, review what it remembers and keeps, and delete
chats; it also keeps a history I'll need to protect and clean
up. Explain that signing in is not the same as giving the AI my
email or files, and that the one sign-in to avoid is a reused
password with no 2-step verification. Then tell me to sign in or
create a free account, start a new chat, and paste this prompt
again. If I choose to stay signed out, tell me what I can't
control (I can't lock it, review it or delete what I share), and
skip to task 3. Signed out, don't offer the fast-track list in
task 5: ask only general types, like "a tablet and a game
console," and tell me to keep my full list on paper.
2. Lock this AI account with a passkey, or 2-step verification. Tell
me to save backup codes on paper, never in this chat.
3. Turn off training, the setting that lets the company use my chats
to improve its AI. In ChatGPT it's under Settings, Data Controls.
4. Check memory and ads. Show me where to read what it remembers
about me, help me delete what it doesn't need, and let me decide
on or off. If my plan shows ads, show me where to turn off ad
personalization, which can use my chats and memory.
OWN IT: Know what we have. Keep this part quick.
5. Ask whether I want the fast track or one question at a time.
Explain in one sentence that our Own IT list is a map of our
family's setup, so I should share ages and types only. Fast
track: give me this fill-in list to answer in one message. No
names, schools, build numbers or serial numbers:
- For each child (by age): phones, tablets, computers, game
consoles, watches and smart speakers they use, and whether each
is family-owned or from school
- Web browsers and email services they use
- AI apps and websites they use (like ChatGPT, Gemini, Copilot,
Meta AI or Character.AI)
- Games and social apps with chat (like Roblox, Minecraft,
Snapchat, Instagram or TikTok)
One at a time: ask about the same things, one kind per message.
School devices are usually managed by the school.
6. Turn my answers into an Own IT list, grouped by child and device,
showing the AI that's likely built into each item (say "check" if
you're not sure). Show it to me, then ask about things families
often forget: a shared family tablet, a grandparent's device, a
smart speaker or TV, an old phone, or an app a friend suggested.
Tell me to save the final list outside any AI tool, and to delete
this chat once I've saved the final summary. Then remind me that
we'll change settings in Secure IT, not now.
7. Help me check each app's minimum age against my kids' ages.
8. Help me decide what my kids may use AI for, and what they should
never share with it.
SECURE IT: Lock the doors. Start with the most important.
9. The AI built into our email, calendar and chat. Ask which email
services we use, like Gmail, Outlook.com, Hotmail, MSN, Yahoo,
AOL or Apple Mail (not our addresses). Explain in two sentences
why it matters: other people, including strangers and hacked
accounts, can send it words, and hidden words can trick it. Tell
me it may already be on, even if nobody turned it on. For my own
account, let me choose Cautious (turn it off), Balanced (only the
feature I use) or Accepting (keep it on, with safety settings).
For my kids' accounts and our shared family email, suggest
Cautious, but let me decide. Keep spelling and grammar on. Tell me
where to look for each service, set images to "ask first" where I
can, and in Google Calendar set "Add invitations to my calendar"
to "Only if the sender is known." If an email has no AI features,
skip it. School email is managed by the school.
10. For each AI app my kids use, one app at a time: turn off training
on chats first, then check memory and history, saved chats and
files, automatic web search, and connections. Explain the rule:
start with no keys, give apps only what they need, and take it
back when they don't.
11. For each child and device, one at a time, help me turn on
parental controls or a supervised account where the app offers
them. Tell me if something can't be changed because it's a school
device.
12. Help me lock my kids' accounts and our shared family accounts
with a passkey or 2-step verification where they can be.
13. Show me where the Stop button is, so I can teach my kids.
PROTECT IT: Share less, check more.
14. Help me write a short family AI agreement we can post at home.
15. Give me a made-up example to practice with my kids: asking a
good question, leaving out private details, and checking the
response together.
16. Help me explain, in kid-friendly words, that AI is a tool, not a
friend, and to tell a grown-up if a chat feels wrong. Also teach
them: if the AI starts acting strange or forgetting things,
start a new chat.
MONITOR IT: Check in and clean up.
17. Give me a short monthly family check-in: what's new, which apps
are installed, new AI in our email (even features nobody turned
on), old chats to delete, and whether our agreement needs an
update.
At the end, give me a summary organized by First, Own IT, Secure IT,
Protect IT and Monitor IT, with our Own IT list and the choices I made
in task 9. List anything I skipped, couldn't find or put on the later
list as still open, not done. Leave out private details. Remind me
that next month we start again at Own IT, and that I can delete this
chat once I've saved the summary.After the conversation
Save the useful result outside AI. Check what the service keeps, remove copies you no longer need, and review any remaining permissions. Deleting a chat may leave separate memories, files, or other retained copies.
The author develops commercial security tools in this area. These resources are tool-neutral and recommend no product. Product names, when used, are examples rather than endorsements. Own IT, Secure IT, and Protect IT draw on the 2019 Cybersecurity Awareness Month theme; Monitor IT extends the cycle into a repeatable personal habit. General educational information, not a scan, compliance assessment, or guarantee of safety. CC BY 4.0.