Apps
ChatGPT
Medium riskAn AI chatbot that can be helpful for learning but also blurs the line between support, shortcut, and authority.
Start here — 3 things to do today
- 1
Turn off training usage
Settings → Data Controls → Improve the model → OFF so conversations are not used to improve the model.
- 2
Disable Memory
Settings → Personalization → Memory → OFF if you want fewer saved details across chats.
- 3
Link a parent account
Settings → Parental controls → Add family member to link your account to your teen's (13-18). This unlocks quiet hours, content protections, and distress alerts (launched September 2025).
- 4
Agree on school rules first
Check your child's school AI policy and write simple family rules for homework, citations, and when AI help is allowed.
What parents worry about most
The goal with ChatGPT is usually honest use, not blocking. Children need rules about what the tool can help with and what still has to be their own work.
The one thing to do right now
A clear family rule about homework and fact-checking is usually more effective than trying to ban the tool entirely.
New since 2025
ChatGPT now has real parental controls: link a teen account in Settings → Parental controls to apply content protections they cannot disable, set quiet hours, and receive alerts if signs of distress are detected.
Risk level
Age rating
13+ official
Users
Massive student and adult user base
Platform
Web, iOS, Android
Age recommendation
Best used with clear rules about homework and fact-checking
None
Social exposure
High
Academic risk
Moderate
Content risk
Emerging
Emotional dependency
Warning signs
Warning signs to know
Jailbreak prompts to bypass safeguards
HighChildren can find or share prompt tricks that try to unlock unsafe, biased, or age-inappropriate responses.
Academic cheating
HighThe tool can write essays, solve homework, and produce polished answers faster than a parent or teacher may notice.
Emotional over-reliance
MediumSome children may begin to use the chatbot as a substitute for real support, feedback, or social contact.
Confident misinformation
MediumChatGPT can sound certain while still being wrong, which is risky when a child treats it like a search engine plus tutor.
Conversation data retention
LowIf settings stay at the default, children may share personal details without understanding how their conversations are stored.
Step-by-step guide
Complete step-by-step guide
- 1
Turn off training usage
Settings → Data Controls → Improve the model → OFF so conversations are not used to improve the model.
- 2
Disable Memory
Settings → Personalization → Memory → OFF if you want fewer saved details across chats.
- 3
Agree on school rules first
Check your child's school AI policy and write simple family rules for homework, citations, and when AI help is allowed.
- 4
Teach that the model can make things up
Use a few test questions together and verify the answers so your child sees that fluent does not always mean correct.
- 5
Set up the new parental controls
OpenAI launched native parental controls in September 2025. In ChatGPT go to Settings → Parental controls → Add family member and send your teen (13-18) an invite they accept. Once linked, content protections (less graphic content, no sexual/violent roleplay) turn on automatically and your teen cannot switch them off. You can also set Quiet hours, turn off Voice mode, Memory, and image generation, and opt out of model training.
- 6
Understand the distress-alert system
With a linked account, if OpenAI's systems detect possible signs of acute distress or self-harm, trained reviewers may notify you by email, text, or push. You can opt out, but for many families it is worth leaving on. There is a parent guide at chatgpt.com/parent-resources.
- 7
Disable optional browsing tools if enabled
Settings → Beta Features and turn off extras you do not want the child using independently.
- 8
Use it together for the first week
Sit with your child for early sessions so they learn how to ask for help, verify answers, and stop when the tool becomes a shortcut.