Parental Controls for ChatGPT, Copilot, Gemini and Claude: What UAE Schools Allow, and How to Set Each One Up

What UAE Schools Actually Allow

This is the part most parenting guides miss, and it is the part that decides what your child is permitted to do tonight.

In February 2026 the UAE Ministry of Education issued Safe and Responsible Use of Artificial Intelligence in Classrooms, a manual circulated to schools nationwide containing 25 prohibitions on generative AI. It names four approved platforms — ChatGPT, Copilot, Gemini and Claude — and bans everything else from the classroom.

Also banned under the manual:

• Using AI during exams or any formal assessment
• Passing AI work off as your own — submitting AI-generated research or reports without disclosure or teacher approval
• Uploading personal data into AI systems — names, photos, audio, video or identity details
• Unapproved platforms in the classroom, and using a VPN to get around school filters
• Creating external accounts for students without permission
• Deepfakes and impersonation, and any material containing violence, hate speech or misinformation, or contradicting UAE values

Students and teachers are both required to "verify the accuracy, relevance and alignment of AI-generated content with approved curricula and learning objectives" before it goes into a lesson or an assignment. The Ministry's framing throughout is that AI is a support tool, not a substitute for a teacher or for a student's own thinking.

And from August 2026, AI is a mandatory subject from kindergarten to Grade 12 in UAE government schools. Your child is going to be taught this. The question is what happens on the devices at home.

Why This Matters Right Now

Schools in Dubai, London and New York run on Microsoft 365 and Google Workspace — which means Copilot and Gemini are inside every assignment your child opens on their school device, whether either of you realises it or not. The UAE added AI to the national curriculum in 2024 and made it mandatory in 2026.

The documented risk In 2025, US parents sued OpenAI after their teenage son died by suicide, alleging ChatGPT had reinforced his harmful thoughts rather than redirecting them. That lawsuit is a direct reason the parental controls in this guide exist — OpenAI shipped them weeks later. A study in Psychiatric Services found that several leading AI chatbots gave inadequate responses when users expressed signs of self-harm risk.

And the everyday risk, which is duller and far more likely to affect your child. Stanford's Graduate School of Education surveyed close to 30,000 high school students and found 32.8% had used AI as an unauthorised aid — almost exactly the same share who admitted copying someone else's homework. Meanwhile the citations these tools produce are frequently invented: one audit of GPT-4o output found 19.9% of citations were entirely fabricated, and of the ones that were real, 45.4% contained errors. A child who hands in AI-written work in a UAE school is breaking the Ministry's disclosure rule and is odds-on to be citing sources that do not exist.

We go through the evidence in detail, risk by risk, in the companion piece: Is ChatGPT safe for children? What the research actually shows. This guide is the practical half — what to switch on, and where.

1. ChatGPT Parental Controls

OpenAI launched parental controls on 29 September 2025, after sustained pressure from child safety advocates and the lawsuit described above. You link your account to your teen's, and from that point you manage their settings from a control page inside your own account. Either side can send the invite — and if your teen later unlinks, you are notified.

Two things have changed since, and both matter.

What you can control once your accounts are linked:

• Quiet hours — blocks ChatGPT completely during the hours you set
• Voice mode — turn it off to remove voice conversations entirely
• Memory — turn it off and ChatGPT won't save or reuse your teen's personal details
• Image generation — turn it off to remove the ability to create or edit images
• Model training — opt your teen's conversations out of being used to improve OpenAI's models
• Study Mode — new. Switch it on here and it is on by default every time your teen starts a new chat, which is the single most useful toggle on this page for a parent worried about homework

Protections that apply automatically, which your teen cannot switch off: Once accounts are linked, the teen account gets reduced graphic content, viral challenges, sexual, romantic or violent roleplay, and extreme beauty ideals. A parent may turn these off; the teen cannot. This is the part most guides miss — linking accounts does something even if you never open a single setting.

How to set it up:

1. Sign into your ChatGPT account — not your teen's — then open Settings
2. Select Parental Controls and send an invite to your teen by email or phone number
3. They must accept before anything is linked
4. Open their control page, set quiet hours, turn off memory, opt out of model training, and switch on Study Mode
5. Confirm safety notifications are on and pointed at a phone you actually check

Safety notifications — and their deliberate limits. If OpenAI's systems detect signs a teen may be about to harm themselves, a trained human team reviews it and you are contacted by email, text and push. On 13 July 2026 this expanded to more situations, including when a linked teen's account is banned for violent activity. OpenAI is explicit that these alerts are meant to be narrow: they are not triggered by fictional writing, gaming, news discussion, political discussion, general anger or abstract questions. Do not expect a running commentary on your child's chats — you will not get one, and you will not see their conversation history either.

What ChatGPT for Teens actually changes. It is a distinct experience, not just a filter. It brings Study Mode, responsible homework reminders that spot a teen trying to shortcut an assignment and steer them back to step-by-step work, quizzes, learning visualisations, and Study Hours — times when Study Mode switches on by itself. On the safety side, OpenAI's under-18 model rules go beyond blocking romantic roleplay: ChatGPT is not supposed to use romantic language, encourage emotional dependence, or imply it has feelings or consciousness. There are sensitive-image upload reminders, break reminders, and added parent notifications covering eating disorders.

For a UAE parent, the homework reminders are the relevant bit. They push in the same direction as the Ministry's rule that AI must support learning rather than replace it.

2. Microsoft Copilot Parental Controls

Microsoft's rule is that "you must be at least 13 years old to use Copilot, and in some countries, the minimum age may be higher." Copilot is not just an app your child chooses to download — it is built into Word, Excel, PowerPoint, Outlook and Teams. If their school runs Microsoft 365, it is already inside every assignment they open.

What you can actually control via Microsoft Family Safety:

• Block Copilot entirely — both the app and the browser version
• Set time limits on when it can be used
• Switch off individual features without blocking the whole thing — image generation, voice mode, group chat and AI companions can each be disabled separately

That third option is the one worth knowing about. You do not have to choose between full access and nothing.

How to set it up:

1. Go to family.microsoft.com or install the Microsoft Family Safety app
2. Create a Family Group and invite your child using their Microsoft account email
3. Under Apps and Games, find Copilot and set the limit or block you want
4. Under Web and Search, add copilot.microsoft.com to blocked sites — blocking the app alone leaves the browser open

One thing on the horizon: on 8 September 2026 Microsoft began rolling out new Windows Age APIs to Insiders, which let apps request a non-identifying age bracket for the signed-in user. It is plumbing for future age-appropriate experiences rather than a control you can use today, and Microsoft's announcement does not add any Copilot-specific parental controls.

3. Google Gemini Parental Controls

Most parents focus on ChatGPT and miss Gemini entirely — yet it is inside Gmail, Google Docs, Search and Workspace. Many schools in the UAE run Google Workspace for Education, which means Gemini is sitting in your child's school email and class documents already.

Google splits this by age. Under 13, a child can use the Gemini app only on a supervised Google account, and you turn that access on or off through Family Link. From 13, they move to a separate teen experience on their own account.

How to set it up:

1. Install Google Family Link and link your child's Google account
2. Open their profile → Controls → Google Apps and set the permission for Gemini Apps
3. Turn on SafeSearch while you are there
4. If you want it gone from the browser too, add gemini.google.com to your blocked sites

4. Claude — 18+ Only, and Enforced

Claude is the only platform here that requires users to be 18 or older, with no exceptions and no teen tier. Anthropic states plainly: "When we detect signals that you may be under 18, we'll ask you to verify your age before you can continue using Claude." Accounts showing indicators of a minor are automatically disabled, paid subscriptions included.

Getting an account back means verifying through Yoti, an independent age-verification provider, by one of three routes: a selfie for facial age estimation, an ID document such as a passport or Emirates ID, or the Yoti Digital ID app. Anthropic says it receives only a pass or fail — never the document or the image — and that Yoti deletes what it was sent once the check is done.

There are no parental controls because there is no child account to control. If your child is under 18, the correct action is not configuration, it is blocking: add claude.ai to blocked sites in Apple Screen Time or Google Family Link.

And this is where the UAE guidance and the platform disagree. The Ministry lists Claude among four tools approved for classrooms; Anthropic bars everyone under 18. Every school student in the country is under 18. If Claude appears on your child's reading list, that is a question for the school, not something to solve at home.

Which Tool, at Which Age — the UAE Answer

Do This Now — Under 30 Minutes

Frequently Asked Questions