When a group of parents demands that you ban AI, do not answer the ban. Answer the worry underneath it, in writing, within ten school days, and give the group three things: an answer, a concession and a date. A demand is a solution someone has already chosen for a worry you have not yet heard. Leaders who reply to the solution end up in a debate about AI. Leaders who reply to the worry end up in a conversation about their school.
That is my suggested way of handling it, not a research finding. What follows is the evidence for why the worry and the demand come apart, a framework you can use this week, and a reply you can adapt.
What are parents actually asking for when they demand a ban?
Mostly they are asking for guardrails, and "ban" is the word people reach for when they cannot see any. The best recent evidence comes with a caveat about who was asked. The Century Foundation, commissioning Morning Consult, surveyed 2,002 U.S. registered voters online between May 21 and 24, 2026, and published the results on August 25, 2026. Of those voters, 85% were concerned about students using AI to complete assignments without learning the material, and 84% were worried about private companies collecting and profiting from students' personal data. Yet when offered a choice, 49% said classroom technology and AI should be kept to a minimum with clear government rules, 28% said schools should embrace technology and AI proven to boost learning, with rules, and 23% said schools should use them however they judge best. Seventy-seven percent, in other words, want rules. Just under half want the minimum.
A similar pattern appears in the parent surveys. Chalkbeat's Lily Altavena, reporting on September 14, 2026, cited a National Parents Union survey from May 2026 in which 53% of parents said it was important for students to learn to use AI tools for future career opportunities, and in which only 1% of the parents who supported screen-time limits wanted zero screen time at school. Chalkbeat did not report the survey's sample size or sponsor, so treat the figures as indicative.
Neither survey tells you what your parents think. Voters are not parents, national samples are not your gate, and a poll question is not a petition. But together they suggest something worth testing in your own meeting: the parents at your door may want a ban less than they want an answer to one of three questions. What is my child's data doing? Is my child still learning to think? Who is accountable if this goes wrong?
Why does the decision so often leave the principal's hands?
The question gets decided wherever the group is heard first, and if the school does not hear it, someone further up the chain will. In the fall of 2026 that happened at the largest scale. Axios reported on September 7, 2026 that New York City's public schools blocked AI use for roughly 600,000 students from pre-K through eighth grade, and that Los Angeles Unified announced a districtwide moratorium on AI use on school devices for 2026-27 hours later. K-12 Dive's Anna Merod, writing on September 17, 2026, described the push in both cities as coming from grassroots advocacy groups organized by parents, community members and teachers.
I cannot tell you what each principal in those districts did, and the reporting does not say. The point I take from it is an interpretation: an organized group that is not answered at the school will look for a body that can. A principal's reply is the last easy chance to keep the question local, specific and reversible. After that, it becomes a district question with a district-sized answer.
There is a second pressure inside your own building. In an NPR/Ipsos poll of 545 K-12 teachers, published June 5, 2026, about half of teachers said their school had not offered AI guidance or that they were unsure what it was. If the parents come with questions and your own staff cannot answer them in the same words, the first job is internal. Settle what staff will say before you reply to anyone.
What is the Ten-Day Reply?
The Ten-Day Reply is the response I suggest a school leader owes any organized group of parents within ten school days: an answer, a concession and a date. The answer says, in plain words, what AI does and does not do in the school today. The concession names one thing the school will change now, whatever it eventually decides about the larger policy. The date says when the decision will be made, by whom, and how the parents can be heard before it is.
Two lines go with it. A demand is a solution someone has already chosen for a worry you have not yet heard. And silence is also a reply, and parents read it as the answer.
Here is how each part works.
The answer. Write what is true today, even where the truth is unflattering. Which AI tools do staff use, and for what? Which tools can students reach, and what do those tools do with student data? What will students still do without AI this term? If you do not know, say so, and say who is finding out and by when. An answer that sounds confident and turns out to be wrong ends the conversation badly; an answer that says "we do not yet know, and the assistant principal will report by the 28th" keeps it going.
The concession. Give something real that asks little of the school in principle and means a good deal to the parents. The test for a good concession is that it is right under every outcome of the larger decision. Publishing the list of AI tools students touch, and what data each one holds, passes the test: it is useful whether you ban, pause or permit. So does naming one person parents can contact, or committing to show parents any new student-facing tool before it launches. If your staff data rules are not written down, the rule on putting student data into a chatbot is a concession you can finish in a week. A pledge such as "AI will never be used in this school" fails the test, because you cannot keep it and everyone will eventually learn that.
The date. Say when the decision will be made, who makes it, and how parents can weigh in first. A principal who cannot decide alone should say so, and name the board, the governors or the district, and the meeting. A group with a date can wait. A group without one escalates.
What should the first meeting with the group look like?
Listen before you defend anything, and ask five questions in this order. The first meeting is not where you win the argument. It is where you find out which of the worries is the real one.
- What have you seen or heard that prompted this? (Ask for the specific thing: an assignment, a tool, a news story, something a child said.)
- Which of your worries matters most: data, learning, or accountability?
- What would you need to see by the end of this school year to feel differently?
- What do you want us to stop, start or tell you?
- Who else shares this view, and who in the parent community sees it differently?
Take notes in front of them and read the worries back at the end. Do not promise a decision in the room. Do promise the date by which they will get the reply.
The fifth question matters more than it looks. An organized group is rarely the whole parent body. The Chalkbeat reporting quotes National Parents Union president Keri Rodrigues warning against a "knee-jerk conversation around this whole tech-lash thing," and the same piece quotes parent advocate Maya Martin Cadogan saying that many children in the communities her parents come from have no device at home. A reply that answers only the loudest group risks overriding the parents who see it differently.
When is the right answer yes?
Yes is the right answer when you can say what you are protecting, why you cannot protect it another way yet, and what you would need to see before lifting the restriction. That is the test from Should Schools Ban AI?, and it applies here unchanged. A restricted period with a named end date and named work to do in the meantime is a perfectly good Ten-Day Reply. A ban with no end date and no reasons is not an answer. It is a mood with a policy number.
The Ten-Day Reply does not tell you what to decide. It tells you what you owe people while you decide. Some groups will be right. If, once you have heard the worry, you find that students' data is going somewhere you cannot account for, then the right concession is to pause that tool until you can. Doing it in the reply, rather than after a district vote, is what keeps the decision yours.
The date in your reply should also feed your governing body. If you are not sure what to report upward, the guide to what to report to governors about AI sets out what a board needs to see, and a short school AI policy gives the reply somewhere to land.
The Mistake I See Most Often
The mistake I see most often is a reply that answers the demand rather than the worry. Across the leadership teams I work with, the first response tends to take one of two shapes. The first defends: a paragraph on the school's careful use of AI and how much teachers value it. The second concedes everything: an immediate promise to review whether AI should be used at all. Both answer the question "ban or no ban," which is the question the group chose. Neither answers the question underneath it, which is why the group is worried, and so both leave the parents where they started, only now with a written record.
I have taught in the classroom, worked as an assistant headteacher, and now serve as a school trustee, so I have seen this from the staffroom and from the board table. From both seats my view is the same, and it is a view rather than a finding: a reply that is specific, dated and short settles more than one that is long and general. The letter-home guide has the Kitchen Table Test, which is a good check on the finished reply.
What does a Ten-Day Reply look like?
It is a one-page letter that does the three jobs in the group's own language. Here is a template, followed by a hypothetical example. Both are adaptable, and neither is a prescribed text.
Thank you for raising this with us, and for meeting with us on [date]. We heard three worries: [data / learning / accountability, in their words].
Where we are today. [Two or three plain sentences: which AI tools staff use and for what; which tools students can use and what data they hold; what students still do without AI this term. Name anything we do not yet know and who is finding it out by when.]
What we are changing now. [One concrete thing: for example, publishing the list of AI tools students use and the data each one holds, by date; or naming one contact for questions.]
What happens next. [Who decides, when, and how you can be heard before then: for example, "The board will take this at its meeting on [date]. We will send you the draft on [date], and you can reply to [name] by [date]."]
A hypothetical example, to show the shape. Suppose about forty parents at a K-8 school write to the principal asking for a ban on student AI use. After the meeting, three worries emerge: that a homework app is sending children's work to a company the parents have not heard of, that children are skipping the writing, and that nobody can say who is responsible. The principal's reply states which apps are in use, says the school does not yet know where one app's data goes and that the assistant principal will report by the 28th, commits to publishing the app list by Friday, and names the board meeting at which the group can speak. Nothing in that reply says "ban" or "no ban," and the group has what it came for.
What should you do this week?
Do four things before the group asks twice. Check what staff will say, because the answer cannot be better than what your own teachers can explain. Write down the tools students touch and the data each one holds, even if the list is rough. Choose the concession you can make that is right under every outcome. And set the date: put the decision on a board or governors' agenda before anyone has to ask you to.
If your leadership team is working through this, it is the kind of problem I support through AI strategy work with schools, where we turn a demand like this into a short, dated plan the whole team can stand behind. If you want the next piece as it appears, the newsletter is the place.
Sources and further reading
- The Century Foundation (Julie Margetta Morgan, Halley Potter and Alex Edwards), Americans Are United Against the Tech Takeover of Public Schools, August 25, 2026. Survey by Morning Consult, May 21 to 24, 2026, 2,002 U.S. registered voters, margin of error plus or minus 2 percentage points.
- Chalkbeat (Lily Altavena), Parents worry about ed tech. Most still see a place for it, polls suggest, September 14, 2026.
- K-12 Dive (Anna Merod), A summer of AI bans in schools, September 17, 2026.
- Axios (Josephine Walker), Parents, students and educators are revolting against AI in classrooms, September 7, 2026.
- NPR, Most K-12 teachers say AI's impact on education will eclipse the internet or computers, June 5, 2026. NPR/Ipsos poll of 545 K-12 teachers.
Dan Fitzpatrick is the founder of The AI Educator, a Forbes contributor and international keynote speaker, and a school trustee who has trained more than 150,000 educators across 30+ countries.
