Is ChatGPT COPPA Compliant?

OpenAI's own Terms of Use set 13 as the minimum age for a ChatGPT account, and the company offers no separate verifiable-parental-consent pathway that would let a child younger than that use the consumer product legally. That single fact answers most of the COPPA question before any contract language even enters the picture: unlike a district-managed tool built around a school-consent mechanism, consumer ChatGPT simply excludes under-13 users rather than providing a compliant way to include them.

Quick Answer: No version of consumer ChatGPT (Free, Plus, or Pro) is COPPA compliant for a child under 13, because OpenAI's Terms of Use prohibit that age group from holding an account at all, and OpenAI has not built a school-as-agent consent pathway comparable to what Google offers for Workspace for Education. Users 13 to 17 need parental consent under OpenAI's own terms. ChatGPT for Teachers, OpenAI's free K-12 tier, is built for teacher use and doesn't currently offer student accounts, so it doesn't change the under-13 analysis. This article covers COPPA specifically, a different federal law with a different test than FERPA, which a companion guide already covers in depth.

This is a COPPA-specific analysis, distinct from the FERPA question Is ChatGPT FERPA Compliant? What Schools Need to Know already covers in detail, including OpenAI's four-tier product structure and its education-terms timeline. Cross-reference that guide for the FERPA-specific legal test; this article stays narrowly on COPPA, verifiable parental consent, and OpenAI's actual stated minimum age.

COPPA Is a Different Law Than FERPA, With a Different Trigger

COPPA regulates the vendor directly and applies specifically to children under 13, while FERPA regulates the school and applies to a much broader "student" population regardless of age. Confusing the two is the single most common mistake a district makes when researching ChatGPT's compliance posture.

The Children's Online Privacy Protection Act, enforced by the Federal Trade Commission under 16 CFR Part 312, requires an operator to give clear notice of what it collects from a child under 13 and to obtain verifiable parental consent before collecting that data. It has nothing to do with the U.S. Department of Education, and nothing to do with whether a 16-year-old's data gets shared with a tutoring vendor, which is squarely a FERPA question instead. AI, FERPA & COPPA: The Complete School Data-Privacy Guide (2026) walks through the full side-by-side comparison of both laws.

OpenAI's Actual Minimum Age: 13, Not 18

ChatGPT's consumer age floor sits at 13, with parental consent required for users 13 through 17, according to OpenAI's own Terms of Use. That's a meaningfully different design choice than Anthropic's approach with Claude, which sets its consumer floor at 18 and offers no under-18 pathway on the general chat product at all.

One sentence takeaway: OpenAI's answer to "what about kids under 13" is exclusion, not a consent mechanism, which is a structurally different approach than Google built for Workspace for Education.

What Changed for Teens in Late 2025

OpenAI introduced a supervised teen mode with automatic content protections for linked 13-to-17 accounts in late 2025, including reduced graphic content, restrictions on romantic or sexual roleplay, and limits on viral-challenge content, according to OpenAI's own Model Spec update announcement. A parent can turn some of these protections off; a teen using a linked account cannot. That update strengthens the teen experience specifically. It does not touch the under-13 question, since a 13-to-17-year-old is a different population than the one COPPA is built to protect.

Why Consumer ChatGPT Doesn't Satisfy COPPA for a Child Under 13

Even setting aside OpenAI's own age-floor rule, consumer ChatGPT's data practices don't fit the narrow school-consent pathway COPPA otherwise allows. This is the part of the analysis that trips up a district assuming "our teacher's account, so it's covered."

The School-Consent Test, Applied to Free and Plus Accounts

Longstanding FTC guidance lets a school consent to data collection on a family's behalf, but only when an operator collects data solely for the use and benefit of the school, for a school-authorized purpose, with no other use layered on top. Free and Plus consumer ChatGPT accounts don't meet that bar on their own: OpenAI's consumer terms allow prompts to be used for product improvement and, unless a user opts out, for model training, a dual purpose that isn't "solely for the benefit of the school." That gap exists regardless of whether a teacher's personal account has toggled training off, because COPPA's school-consent pathway is about the operator's structural data use, not one individual's personal setting.

There's No Under-13 Consent Flow to Point To

Unlike Google, which built explicit school-as-agent consent language into its Workspace for Education Terms of Service specifically to enable Core Services use by students under 13, OpenAI's consumer terms simply set the age floor at 13 and stop there. There is no parallel document a district can point to that says "here is how a school consents on behalf of a family for a ChatGPT account held by an 11-year-old," because OpenAI's terms don't contemplate that account existing in the first place.

The practical result: a school that lets a fourth grader use Free or Plus ChatGPT, even under a teacher's supervision on a shared device, is operating outside both OpenAI's own Terms of Use and COPPA's actual requirements, since no verifiable parental consent mechanism sits behind that access at all.

Does ChatGPT for Teachers Change the Under-13 Analysis? No, and Here's Why

ChatGPT for Teachers, free for verified U.S. K-12 educators through June 2027, is explicitly a teacher-facing workspace, not a student-facing one, and that design choice is exactly why it doesn't touch the under-13 COPPA question.

• The product doesn't currently offer student accounts. OpenAI's own help documentation describes ChatGPT for Teachers as built around the individual educator's workspace, with student-facing features described as a possible future addition rather than a current capability.
• Content shared inside ChatGPT for Teachers is not used to train models by default, and the workspace carries FERPA-aligned terms and a student data privacy agreement a district can review, addressing the FERPA question this guide's companion article covers in depth.
• None of that changes who is actually typing into the product. Since a teacher, not a student, holds the account and drives the interaction, COPPA's under-13 trigger, direct collection of personal information from a child, doesn't activate the same way it would if a nine-year-old held their own login.

For a school specifically weighing whether a young student can interact with ChatGPT at all, the honest answer today is that any legitimate access has to route through an adult using the tool on the student's behalf, not a student-held account, regardless of which ChatGPT tier the district has licensed.

ChatGPT Edu and Enterprise: Built for Institutional Scale, Not Younger Students

ChatGPT Edu and ChatGPT Enterprise add admin-managed provisioning, SSO, and custom retention windows, but neither product was designed around, or marketed toward, students under 13. OpenAI's own materials position ChatGPT Edu for universities and larger, typically secondary-level, school systems.

A district that has licensed ChatGPT Edu for a middle or high school still needs to separately confirm, in writing, whether any student provisioned under that agreement is 13 or older, since the underlying 13-plus age floor doesn't disappear just because the district, rather than an individual, is managing the account.

Common Misunderstandings About ChatGPT and COPPA

A handful of assumptions come up repeatedly when a district first researches whether ChatGPT is COPPA compliant, and most of them borrow reasoning from FERPA that doesn't actually transfer.

• "OpenAI has a signed DPA with our district, so COPPA is covered too." A FERPA-relevant Data Processing Agreement addresses the school-official disclosure test. It doesn't, by itself, create a verifiable-parental-consent mechanism for a child under 13, since that's a separate legal requirement with its own separate test.
• "ChatGPT for Teachers is free for K-12, so it must be designed for K-12 students." It's designed for K-12 educators. The product's entire value proposition is a teacher-managed workspace; treating it as a student on-ramp misreads what OpenAI actually built and shipped.
• "If our older students can legally use ChatGPT, our younger ones probably can too, just with more supervision." Supervision doesn't substitute for the account-holder's age. OpenAI's 13-plus floor is a hard eligibility rule for the account itself, not a risk level a school can offset with more adult oversight.
• "COPPA and FERPA are basically asking the same question about ChatGPT." They're not. AI, FERPA & COPPA: The Complete School Data-Privacy Guide (2026) lays out why the two tests diverge, and applying FERPA logic to a COPPA question is exactly how a district ends up with a policy gap nobody noticed.

How This Compares Across the Three Major AI Vendors Schools Actually Ask About

Placed side by side, OpenAI, Anthropic, and Google take three genuinely different approaches to the same under-13 problem, and none of them is simply "more compliant" than the others in the abstract. Each reflects a different product philosophy as much as a different legal reading.

OpenAI's approach is exclusion by age floor: set the line at 13, and don't build a mechanism for anyone younger. Anthropic goes further in the same direction, setting its consumer floor at 18 and pushing any minors-facing use onto API-integrating organizations that build their own age verification and content moderation, as Is Claude Safe for Students? covers in more depth. Google took the opposite path for its Core Services specifically: rather than excluding under-13 students, it built a contractual school-as-agent consent mechanism designed to include them lawfully, which Is Google Classroom COPPA Compliant? examines in detail.

None of these approaches is inherently safer or riskier than the others from a K-9 teacher's vantage point; they simply mean the practical answer to "can my students use this" depends entirely on which vendor, and which specific product tier from that vendor, a district is actually asking about.

A Practical Checklist for the Under-13 ChatGPT Question

Work through these before assuming any ChatGPT tier is appropriate for a specific age band:

1. Confirm the actual age of every student who might get provisioned access, not just the grade band's typical age, since a held-back or accelerated student can sit outside the expected range.
2. Check which ChatGPT product is actually in use. Free and Plus consumer accounts carry no district-level contract and no under-13 pathway at all.
3. If the population includes students under 13, treat direct student use as off the table entirely, and plan for teacher-mediated use instead, the same posture this guide's companion article recommends for Claude.
4. Verify whether ChatGPT for Teachers is actually the product in play, since its teacher-only design sidesteps the under-13 trigger by construction, not by a specific consent mechanism.
5. Separately confirm your FERPA posture for any student 13 or older interacting with a district-licensed tier, since FERPA's school-official test still applies once COPPA's under-13 trigger no longer does.
6. Document the decision in writing, not as an assumption a specific teacher made once, since a compliance review needs a paper trail, not institutional memory.

A Grade 4 Classroom Scenario, Walked Through

Say you teach Grade 4 and a student mentions using "the chatbot my older sibling uses for homework" and asks whether the class can try something similar for a research project. Here's how the pieces above actually apply.

• A Grade 4 student is almost certainly under 13, which puts this squarely in the population OpenAI's own terms exclude from holding a ChatGPT account, consumer or otherwise.
• You, as the teacher, could use ChatGPT for Teachers yourself to draft research-project scaffolding or differentiated prompts, since that keeps the account holder an adult and sidesteps the under-13 question entirely.
• Letting the student type directly into any ChatGPT tier, even on your device and even supervised, isn't a COPPA-compliant workaround, since the account itself, not just the supervision, is what the age floor governs.
• If your building wants students that age generating their own content with AI assistance, a platform built around teacher-configured class profiles rather than individual student chat accounts sidesteps the age-floor question differently: EduGenius can generate a research-project worksheet or guided-inquiry handout from a class profile set to Grade 4, with no student ever needing to hold an AI account of their own.

Related Reading

• Is ChatGPT FERPA Compliant? What Schools Need to Know
• AI, FERPA & COPPA: The Complete School Data-Privacy Guide (2026)
• Diffit and Student Data Privacy: What Teachers Should Know
• Brisk Teaching and Student Data Privacy: What Teachers Should Know
• Is Brisk Teaching Safe for Students?
• Is Claude Safe for Students?
• SchoolAI vs Khanmigo: Which Is Better for Teachers?