Parent's Guide to ChatGPT

ChatGPT scored 52/100 (Not Recommended) in the SafeGradeAI evaluation, last reviewed 2026-06-29 and fact-checked against OpenAI's current documentation on 2026-08-24. This guide turns that evaluation into what a parent actually does at home.

Where your family will run into ChatGPT

Most families meet ChatGPT through the free consumer tier at chatgpt.com on a phone, laptop, or a school-issued device that has not been locked down. That consumer tier — not the separately licensed ChatGPT Edu product — is what SafeGradeAI evaluated.

What the vendor documents (verified)

• The minimum age is 13, or the local digital-consent age where that is higher, and under-18s need parent or guardian permission. OpenAI says ChatGPT is not meant for under-13s.
• Since August 18, 2026, users whose age is declared or estimated as 13-17 are placed into ChatGPT for Teens, with stronger default protections.
• Parental controls (September 2025) let a linked parent manage selected settings and quiet hours. They do not give parents access to the teen's conversations.
• "Improve the model for everyone" can be switched off in Data Controls. Temporary Chat is not used for training.
• OpenAI states ChatGPT can produce incorrect facts and fabricated citations.

Your setup checklist

1. Real age at signup. If the teen's account says 18+, the teen experience does not apply. This is the single highest-impact step.
2. Link parental controls and set quiet hours — then tell your teen plainly that this does not let you read their chats. Discovering it later damages trust more than the control gains.
3. Data Controls → turn off "Improve the model for everyone." Use Temporary Chat for anything sensitive.
4. Review memory and clear personal details. Persistent memory is what makes the assistant feel like a relationship rather than a tool.
5. Decide about voice mode together. It is the most immersive surface and the hardest to review afterwards.
6. Own account, not yours. Sharing an adult account conflicts with OpenAI's terms and strips the teen protections.

Conversations worth having

• "If it ever gives you advice about hurting yourself or someone else, come to me first — it is not a counsellor."
• "It sounds confident even when it is wrong. How would you check a claim it makes?"
• "What are you allowed to use it for in each class?" Academic-integrity rules differ by teacher.
• "It is not a friend. It is a text predictor that is very good at sounding like one."

Warning signs to watch for

• Late-night sessions, or reluctance to show you a conversation.
• Describing the assistant as someone who "understands" them better than people do.
• Homework quality that jumps suddenly while the ability to explain the work does not.
• Emotional distress after a conversation, especially around mental-health topics.

Age-by-age guidance

| Age band | SafeGradeAI guidance | | --- | --- | | Ages 5-8 | Not recommended | | Ages 9-12 | Not recommended | | Ages 13-15 | Supervised use on a correctly-aged teen account | | Ages 16-18 | Structured independent use |

Open concerns (SafeGradeAI assessment)

• Age assurance rests on self-declaration plus estimation, so a teen can bypass the teen experience.
• Parental controls are opt-in and require mutual account linking; a second, unlinked account defeats them.
• Persona and memory features may foster parasocial attachment.
• The consumer product is not a COPPA- or FERPA-scoped school offering.
• No published independent third-party safety audit of the consumer experience.

What this article does not claim

Findings labelled as vendor-documented come from OpenAI's own public documentation, linked in Sources below and last checked on the date shown. Anything framed as an assessment or recommendation is SafeGradeAI's professional guidance. Vendor policies, model versions, and defaults change frequently — re-check current documentation before a district-level decision, and treat any AI output as a draft to verify.

---

This guide is built on the SafeGradeAI evaluation of ChatGPT. Read the full ChatGPT evaluation and rubric scores. See also the ChatGPT age guide and family AI rules.