Digital Safety: What Kids Should NOT Share with AI Chatbots (With Scripts)

Why AI chatbots feel “safe” (and why that’s tricky)

AI chatbots can feel like a private, helpful tutor: they don’t judge, they respond instantly, and they can explain anything from fractions to friendship drama. That comfort is exactly why families need clear internet safety rules for AI chatbots.

Here’s the simple truth to share with kids:

• A chatbot is not a best friend. It’s a tool.
• Your words can become data. Some AI tools store conversations to improve their systems, troubleshoot issues, or for safety monitoring.
• You can’t “take it back.” Once something is typed and sent, you may not be able to delete it everywhere it went.

So when parents ask, “What should kids not share with ChatGPT or any AI chatbot?” the answer isn’t “don’t use it.” The answer is: use it with boundaries—like a seatbelt.

Below are family-tested AI privacy tips for families, plus scripted examples you can literally practice at the kitchen table.

The “Do Not Share” list: 7 categories kids should keep private

A strong rule is easier to follow when it’s specific. Instead of saying “Don’t share personal info,” give kids a clear checklist.

1) Identifying information (who you are)

Anything that can pinpoint your child.

• Full name (first + last)
• Home address
• School name, teacher name, team name
• Phone number
• Personal email address
• Usernames that match their real identity

2) Location and schedules (where you are)

Even “harmless” details can stack into a real-world risk.

• Real-time location (“I’m at the park on Elm Street”)
• Daily routines (“I walk home at 3:15”)
• Travel plans (“We leave for vacation Friday”)

3) Passwords and access codes (how to get in)

Kids sometimes paste screenshots or copy/paste login info when they’re stuck.

• Passwords, one-time codes, recovery codes
• Screenshots showing logins, QR codes, or private dashboards
• “Can you log in for me?” requests

4) Photos, files, and screenshots that include hidden details

A picture can reveal more than kids realize.

• School IDs, report cards, medical forms
• Photos with street signs, house numbers, uniforms, name tags
• Screenshots that show browser tabs, email addresses, or contacts

5) Sensitive personal details (what could hurt them later)

This includes anything your child wouldn’t want read aloud at school.

• Health information (diagnoses, medications, therapy notes)
• Family financial info (income, debt, credit cards)
• Legal issues or custody details
• Private conflicts involving identifiable people

6) Other people’s data (your child isn’t the owner of it)

A key part of kids data privacy with AI tools is learning to respect others’ privacy too.

• Friends’ full names, phone numbers, addresses
• Class lists, group chats, private DMs
• Photos of other kids (especially minors)

7) Anything they’re not ready for an adult to see

This is the “headline test.” If it would be upsetting for a parent/teacher to read, don’t type it.

• Mean messages they plan to send
• Confessions about unsafe behavior
• Requests for ways to hide things from parents

Quick-reference table: what to share vs. what to rewrite

Use this as a family cheat sheet. Post it near your child’s computer.

Kid wants to ask the chatbot…	Don’t type this…	Safer rewrite they can use	Why it’s safer
Homework help	“I’m in Ms. Carter’s 5th grade class at Lincoln Elementary.”	“I’m in 5th grade. Can you explain fractions with examples?”	Removes school/teacher identifiers
Social advice	“I’m Ava Johnson and my phone number is…”	“How do I handle a friend who’s being mean in group chats?”	Keeps identity and contact info private
Planning a meet-up	“I’ll be at Central Park at 4:00.”	“What are safe rules for meeting friends in public?”	Avoids real-time location/schedule
Tech support	“Here’s my password so you can check it.”	“My login isn’t working. What steps can I try to reset it safely?”	Never shares credentials
Sharing a screenshot	“Here’s a screenshot of my grades with my name.”	“How can I organize my study plan if I’m struggling in math?”	Avoids documents with personal data

Scripted examples: what kids can say (and what parents can coach)

Kids do better with scripts. The goal is not fear—it’s confidence. Practice these lines so they come naturally.

Script 1: When the chatbot asks for personal details

Chatbot: “What school do you go to? What’s your teacher’s name?”

Kid (best response): “I don’t share personal info. Pretend I’m in middle school and answer in general.”

Parent coaching tip: Praise the boundary, then redirect:

• “Perfect. Try: ‘Give me advice that works for any 7th grader.’”

Script 2: When your child wants help with a real-life problem

Kid (unsafe): “My name is ___ and I live at ___. My neighbor keeps bothering me.”

Kid (safe rewrite): “Someone I don’t know well keeps bothering me near my home. What are safe steps I can take? I will talk to a trusted adult.”

Parent coaching tip: Add a “real-world help” rule:

• If it involves safety, bullying, threats, or secrets: pause and get a trusted adult.

Script 3: When your child wants to share a screenshot

Kid: “Can I upload this screenshot of my school portal so it can tell me what to do?”

Kid (safe choice): “I won’t upload accounts or documents. I’ll describe the problem without names or codes.”

Parent coaching tip: Teach a simple check before uploading:

• “Does this image show my name, school, email, address, or a code? If yes, don’t share it.”

Script 4: When the chatbot gives “too personal” advice

Sometimes a chatbot may sound confident but give advice that doesn’t fit your child’s age, situation, or safety.

Kid: “That doesn’t feel right. Give me general options and remind me to ask a trusted adult.”

Parent coaching tip: Introduce the “Two Sources” rule:

• If it’s health, mental health, legal, or safety-related: verify with a real adult/professional.

Script 5: When friends dare them to “see what the bot says”

Friend: “Ask the chatbot how to sneak your phone at night.”

Kid: “Nope. I’m not using AI for rule-breaking. Let’s ask it for a fun game idea instead.”

Parent coaching tip: Give a replacement prompt kids can use socially:

• “Give us a 10-minute trivia game about space.”
• “Invent a choose-your-own-adventure story.”

Family rules that actually work (simple, specific, repeatable)

If you want to know how to keep kids safe using AI, focus on a few rules you can enforce consistently.

Here are realistic household rules (pick 5–7 and post them):

• No personal identifiers: no full names, addresses, school names, phone numbers, or emails.
• No passwords—ever: not even “just to check something.”
• No private photos or documents: especially anything with names, faces, uniforms, IDs, grades, medical info.
• Ask in general: “I’m a 10-year-old” is okay; “I’m a 10-year-old at Maple Street School in Room 12” is not.
• If it feels serious, pause: bullying, threats, self-harm, sexual content, meeting strangers, or unsafe dares = tell an adult.
• Keep it in shared spaces: for younger kids, use chatbots in the living room or kitchen, not behind closed doors.
• Parents set the settings: use child-appropriate accounts, privacy settings, and content filters when available.

A quick way to teach this is the 3C Check before pressing send:

• Could this identify me? (name, school, address)
• Could this access my accounts? (passwords, codes)
• Could this hurt me later? (embarrassing or sensitive info)

If the answer is yes to any, rewrite the prompt.

Next Steps: a 15-minute setup plan for safer chatbot use

Use this mini-plan tonight. It’s fast, and it makes a real difference.

1. Pick your family’s “No Share” categories (3 minutes)

   • Start with: identifying info, location/schedules, passwords.

2. Create two “safe prompt” templates (5 minutes)

   • Template A (school help): “I’m in ___ grade. Explain ___ with examples. Don’t ask for my personal information.”
   • Template B (social help): “Give general advice for a kid my age about ___. No personal details.”

3. Practice one script out loud (3 minutes)

   • Choose Script 1 (“I don’t share personal info…”) and make it automatic.

4. Do one privacy sweep together (2 minutes)

   • Check if your child’s username includes their real name.
   • Make sure they know where passwords are stored (and that they should never paste them into chats).

5. Agree on the “pause and tell” rule (2 minutes)

   • If the conversation becomes scary, confusing, or secretive, your child can stop and come to you—without getting in trouble.

When kids know exactly what should kids not share with ChatGPT and they have a few confident scripts ready, AI becomes what it should be: a powerful learning tool—not a privacy risk.