Explain
Complex topics can be explained in simpler words.
INTERACTIVE GUIDE · ABOUT 12 MIN.
AI Safety
Use AI wisely, verify information, and protect personal data.
Outcome: A personalized family plan for safe and responsible use of AI.
For children, teens, and familiesLast reviewed: 2026-07-14
Artificial intelligence can explain, translate, come up with ideas, create images, and assist with many tasks.
But it can also make mistakes, invent information, provide biased answers, or create convincing fakes.
This guide will help you use AI wisely, protect your private information, and recognize when a person, a reliable source, or professional help is more important.
What you'll achieve in the end
After this guide, you will have:
- understood why AI can make mistakes
- clarified which data should not be shared
- discussed rules for homework and studying
- better understood deepfakes and voice clones
- learned about common AI scams
- determined when to ask a person or expert for help
- prepared family rules for using AI
How safely are you using AI already?
Discuss these points together. Open issues are a guide — not a guarantee of safety.
Well Prepared
0–2 open points
0 open items
The result is a guideline and not a guarantee of safety.
1. How AI Can Help
Develop Ideas
AI can help gather ideas, organize them, and plan.
Language
Texts can be translated, summarized, or improved.
Creativity
AI can create images, stories, music ideas, or design suggestions.
Learning
AI can generate practice questions, examples, and alternative explanations.
Organization
AI can prepare lists, plans, templates, and initial drafts.
AI should assist — not completely replace your own thinking.
2. What AI Can't Do Reliably
AI systems can sound confident, even when their answers are wrong.
Guaranteeing Truth
A convincing answer can still be wrong or made up.
Automatically Evaluating Sources
AI can summarize an unreliable source just as convincingly as a reliable one.
Fully Understanding Intentions
Irony, personal conflicts, or complex situations can be misunderstood.
Taking Responsibility
The decision always rests with the person.
Truly Experiencing Emotions
A chatbot can express empathy, but it has no real feelings or personal connections.
Guaranteeing Timeliness
An answer can be outdated or incomplete.
Well-written doesn't automatically mean it's correct.
3. What You Shouldn't Share with AI
Don't Share
- Passwords
- Verification codes
- Full home address
- Private phone numbers
- Bank or card details
- Identification documents
- School or work accounts
- Confidential family information
- Medical records
- Intimate pictures
- Private messages from others
- Data that can clearly identify a child
Check First
- personal photo
- voice
- homework with names
- specific school
- travel plans
- location
- resume
- private conflicts
- company information
- unpublished work
Mostly Uncritical
- general knowledge questions
- fictional examples
- anonymized practice texts
- brainstorming ideas
- public information
- made-up people and scenarios
- general language exercises
Check the privacy and account settings of the service you're using.
Would I be bothered if this information was saved, checked by an employee, or accidentally made public?
AI can support learning. However, submitting an AI-generated work as your own can violate school rules and hinder your learning.
Good Use
- get explanations for difficult terms
- create practice questions
- review an outline
- receive feedback on your own draft
- look for counterarguments
- practice vocabulary
- simplify a text
- collect source ideas
Risky or Inappropriate
- submitting a complete assignment created by someone else as your own
- using made-up sources
- including incorrect quotes
- bypassing exam rules
- entering classmates' personal information
- not labeling AI-generated content when required by the school
Five Steps for Meaningful Learning
- Think for yourself first
- Ask AI specific questions
- Check the answers
- Revise on your own
- Be transparent about your use when necessary
What should you really understand after the task?
School rules vary. Discuss together what AI usage is allowed.
STOP Method
- SIs this statement really true?
- TDoes the answer have credible sources?
- OCheck online with reliable sources
- PConsider different perspectives and possible gaps
- PAsk someone if the decision is important
- Is there a publication date?
- Is the source official or recognized?
- Do multiple independent sources confirm the statement?
- Are specific studies or documents mentioned?
- Do the cited sources actually exist?
- Could the information be outdated?
- Is the statement an opinion, assumption, or verifiable fact?
Weak question: "Is that true?"
A better question is: "Which parts of this answer are uncertain, what sources should I check, and what counterarguments exist?"
AI-generated quotes can be incorrect or may not exist at all.
9. When AI Isn't Enough
Medical
AI can provide general information, but it cannot replace an examination, diagnosis, or personalized treatment.
Legal
Laws vary by country and situation. Important decisions should be reviewed with a qualified professional.
Emergency
In cases of immediate danger, violence, self-harm, suicidal thoughts, or serious illness, a chatbot is not the right help.
In Immediate Danger
Get help right away — a chatbot is not an emergency, crisis, or counseling service.
- USA and Canada: 911
- European Union: 112
13. Your Family Rules for AI
Choose suitable rules or add your own wording.
Your AI Safety Plan
Well prepared
- Well prepared
- —
- Still open
- —
- Privacy
- Do not enter personal data
- Consent for AI images
- Check sources
- Use the STOP method
- School and Learning
- Clarify school AI rules
- Deepfakes
- Agree a code word and callback rule
- Chatbots and Relationships
- A chatbot is not a substitute for real support
- Next Exam
Selected Rules
- AI answers are checked for important topics.
- Personal data is not entered.
- Passwords and confirmation codes stay private.
- Homework is not fully completed by AI.
- Distressing AI outputs may be shown at any time.
- A chatbot does not replace real relationships.
Establishing AI Rules Together
In the family phone contract, you can agree on privacy, homework, consent, and asking for help in a binding way.
Related Resources
Scams
Recognize pressure, verify identities, and protect money, accounts, and personal data.
Digital Terms Explained
From algorithm to two-factor authentication — short, clear, and family-friendly.
Password and Account Security
Strong passphrases, secure recovery, two-factor protection — and a clear plan for emergencies.
Privacy Check-Up
A quick check for profiles, messages, location, and device access — no security guarantee.
Disturbing Content
Close it, don't forward it, seek support — without shame.
Conversation Guides
How to start, what to ask, what to avoid — and how to respond to anger, silence, or shame.
Family Phone Agreement
A shared promise — not a punishment. Rules apply to everyone. Clarity, trust, and safe habits.
Help Lines
International helplines and official resources — filter by country.
This guide offers general educational and organizational support. AI systems, features, and privacy policies may change. The guide does not guarantee safety and is not a substitute for medical, psychological, legal, educational, or emergency advice.