Age, parental controls, and safe use
Review current OpenAI age guidance, what linked parental controls can and cannot do, and a first-session checklist.
Open the ChatGPT safety guideFamily AI safety hub · ages 7–14
AI safety for kids begins with habits, not a perfect setting
Parents should check age rules, limit personal information, supervise younger children, verify important claims, discuss bias and deepfakes, and keep AI from replacing trusted people. Provider controls can help, but no control removes the need for family boundaries.
Adult chooses accessProvider rules come before convenience.
Minimum dataUse fictional or public examples.
Verify important claimsFluent wording is not evidence.
People remain primaryAI is not a friend, therapist, or authority.
Published: August 2, 2026 · Last materially reviewed: August 2, 2026 · AI Education for Kids Editorial Team
Use before every session
The five-step PAUSE routine
PPermissionCheck age rules, adult consent, and whether an account is appropriate.
AAvoid private dataNo names, schools, addresses, schedules, health details, passwords, or private media.
UUnderstand the taskPredict the result and decide what evidence would count.
SSource-checkVerify important facts, quotations, citations, and images.
EExit and explainStop if uncomfortable; tell an adult and explain what the tool did.
Eight risks parents can act on
Name the risk, then choose a family response
Privacy
What can happen
Prompts, files, voices, images, or account details may be stored or used according to provider settings and policies.
Family response
Use minimal, fictional inputs; review data controls; delete unneeded chats and uploads.
Hallucinations
What can happen
A system can invent facts, sources, quotations, or confident explanations.
Family response
Check high-impact claims with primary or trusted sources before reuse.
Bias
What can happen
Outputs can reflect gaps and unfair patterns in data, labels, and design choices.
Family response
Ask who is missing, compare examples, and avoid using AI to judge people.
Deepfakes
What can happen
Synthetic images, audio, and video may imitate real people or events.
Family response
Pause before sharing, find the original source, and check reliable reporting.
Emotional reliance
What can happen
A conversational system may feel caring or authoritative even though it generates responses from patterns.
Family response
Keep personal crises, health decisions, and emotional support with trusted adults and qualified professionals.
Homework integrity
What can happen
Children may submit generated work they cannot explain or violate a teacher’s rules.
Family response
Ask the teacher’s policy, disclose allowed AI help, save the child’s drafts, and require explanation.
Unsafe content
What can happen
Filters can fail, and a child may see advice or material that is frightening, sexual, violent, or dangerous.
Family response
Use age-aware settings, stay nearby, teach immediate exit, and report serious incidents.
Overuse
What can happen
Easy answers can crowd out sleep, movement, reading, practice, and human conversation.
Family response
Set a clear purpose and time boundary, then end with an offline action or discussion.
Family agreement
Seven boundaries children can remember
We agree that AI may help us think, but it does not make decisions for our family.
- We ask an adult before creating an account, uploading a file, or sharing a result publicly.
- We never enter private identity, location, school, health, password, or payment information.
- We use fictional examples when a real person’s details are not necessary.
- We verify important claims and label content made with AI when context requires it.
- We do not use AI to bully, impersonate, deceive, cheat, or judge another person.
- We stop and tell a trusted adult when a response feels unsafe, secretive, manipulative, or upsetting.
- We can take a break at any time. Finishing a prompt is never more important than wellbeing.
Print this page from your browser to keep the PAUSE routine and family agreement near the shared device.
Choose guidance by situationChatGPT Tool choice Home learning
Go deeper where your family needs it
Compare five common AI tools
See current age rules, account expectations, parent controls, learning value, and supervision notes.
Compare AI tools for kidsPlan supervised AI learning
Set goals, protect privacy, document work, and keep adult judgment visible across a home-learning program.
Read the homeschooling guidePrimary sources
Check current provider and regulator guidance
- OpenAI: ChatGPT age guidance
- OpenAI: parental controls and their limits
- Microsoft: Copilot age limits and parental controls
- Google: guide a child’s Gemini experience
- U.S. Federal Trade Commission: children’s privacy guidance
Reviewed August 2, 2026. Provider settings and age rules change. Follow the linked source and the rules shown inside the family’s account before use.
Practice safe AI with a guided first lesson
The free class introduces AI basics, a practical demonstration, hallucinations, privacy, and responsible use with a parent or guardian involved.