AI Bans in Schools: Why Switching AI Off Is Only Step One

AI bans in schools are back, and this time switching AI off is becoming the default setting. Google says its artificial intelligence (AI) tools will be off for pupils’ accounts until a school turns them on. Seattle is blocking student-facing AI from Monday. Two more United States (US) districts are drafting bans by age.

The short version: switching AI off gives schools time, but it does not teach anyone anything. The districts making the most sense are pairing a pause with staff training, AI literacy lessons that need no chatbot, and protected access for pupils who rely on assistive tools.

What changed in the last two weeks

On 3 November, Google announced that Gemini, Gemini Notebook, Gemini in Classroom and Google Search will be “entirely off for K-12 student accounts until school administrators explicitly turn them on”. K-12 is the American term for kindergarten to 12th grade, so in practice this means every primary and secondary pupil account. Google says it will share details with schools “in the coming months” before the change rolls out. It also launched Guided Learning in Google Classroom worldwide. This mode coaches pupils through a problem with questions instead of handing over the answer.

The next day, at a board meeting on 4 November, Seattle Public Schools said it would block Gemini, Minecraft, generative AI tools and Google Search AI summaries for all students from Monday 9 November, according to FOX 13 Seattle. Pupils in grades 3 to 5 will also no longer have a laptop each. The superintendent, Ben Shuldiner, was candid: he said that when he arrived he saw “a district that was overwhelmed by technology and said, ‘Go to it’, and now we are being thoughtful and critical.”

In Maine, the South Portland School Board votes on Monday on a moratorium on AI for pupils from pre-kindergarten to 8th grade for the rest of the school year, reports the Press Herald. Fairfax County in Virginia has gone further on paper. Its draft emerging technology policy, presented on 27 October, would ban AI entirely from kindergarten to 6th grade, allow a few “curriculum-grounded tools” in grades 7 and 8 with parental approval, and open up wider use in high school only once students have shown AI literacy. A vote is expected in February.

England is heading the same way by a different road. In the House of Lords on 26 October, the education minister Baroness Smith of Malvern said the government would consult “this autumn” on safety certification for higher-risk products. That would lead to a published list of child-facing generative AI tools that meet safety criteria, including data protection and safeguarding.

Why AI bans in schools are back on the table

I understand the instinct. Many schools let AI arrive through software updates rather than decisions. A pause gives leaders time to ask the questions they skipped.

The worry about learning is also real. South Portland board member Claire Holman put it simply: pupils “don’t learn to struggle and push through when something gets difficult.” That matches the research we covered in AI and homework, where scores improved with AI help and then fell once the tool was taken away.

But a ban answers one question, which is whether pupils may use a tool in school. It does not answer what they should learn about it, and that second question does not go away when the filter goes on.

A ban decides what pupils cannot use in school. It says nothing about what they should learn, and that is the harder question.

What a ban does not fix

Pupils still have AI at home

Blocking Gemini on a school account does nothing about the phone in a Year 9 pupil’s pocket or the laptop on the kitchen table. If school is the one place AI is never discussed, pupils learn about it from friends and adverts instead. Our piece on AI chatbots and safeguarding found that even the teen settings on popular chatbots missed warning signs that adults would want flagged.

Nobody has worked out enforcement

The South Portland proposal, by the Press Herald’s account, contains no plan for enforcement, and there is concern that pupils will find ways round it. A rule that cannot be checked tends to punish the honest pupil and miss the confident one.

The exemptions are the interesting part

South Portland’s draft would not apply to technology used by students with disabilities or by English language learners. In Fairfax, the chair of the board’s governance committee said the drafts did not yet have “a strong enough accessibility connection”. The uses people most want to keep are the ones that remove a barrier, such as a reading passage simplified to the right level or a letter home translated for a family. Sarah Gay, an English teacher in South Portland, described using AI for both.

What school leaders can do this term

None of this needs a new policy by Friday. It needs decisions made on purpose.

• Treat the Google change as a decision point. If your school uses Google Workspace for Education, someone will soon have to switch AI on for pupils rather than discover it is already on. Decide now who holds that switch and which year groups it covers.
• Map access by age, as Fairfax has done. A secondary school might keep pupil AI off in Years 7 and 8. Year 10 science could then use one teacher-assigned Guided Learning activity, with wider use in the sixth form.
• Teach AI literacy without the tool. The Fairfax draft asks for “grade-appropriate AI literacy, awareness, safety, and human-judgment instruction” without pupils needing the tools banned for their age. A Year 6 class can compare a printed chatbot answer about volcanoes with a textbook page and hunt for the confident mistake. Our guide to AI literacy in schools has more ideas.
• Protect access for pupils who need it. Write the exemption for special educational needs and disabilities (SEND) and for pupils learning English as an additional language (EAL) before the ban, not after a parent complains.
• Train staff first. South Portland’s resolution would require five hours of AI awareness and literacy training for every member of staff this year. That is a modest, sensible floor.

Then tell parents what you have decided and why. Seattle’s announcement came as a parent group rallied outside the meeting, asking among other things for a published list of every app used in classrooms. That is a fair request.

Off by default, on by decision

Where do I land? I would rather see a deliberate on switch than a permanent ban. Google’s own pitch is that Guided Learning supports “productive struggle”, and it says its research shows learning gains. That is the company’s own evidence, so treat it as a claim to test in your classroom, not a settled finding. The certification list promised in England should help with safety. It will not tell a head of maths whether a tool helps Year 9 understand fractions.

So use the pause. Train staff, teach pupils how these systems get things wrong, and switch on one tool for one purpose when you can say what success looks like. More on the wider picture is in our AI in Schools hub.

Frequently asked questions

Is your school switching AI off, keeping it on, or waiting to see? Tell us in the comments what your leadership team has decided, and whether it is working.