reluctant-citizen

education

My kids will spend their whole adult lives working, learning and living alongside artificial intelligence. That isn't a prediction I'm especially excited or scared about. It's just the direction the world is moving. So when I hear about schools that block AI tools outright, my first reaction as a parent is worry. Not that my kids will be harmed by AI, but that they'll be left to figure it out alone, or not at all, while kids elsewhere learn to use it well.

This post is my attempt to think that worry through honestly: what the rest of the world is doing, what the evidence says about the real risks, and what parents can actually do when their school seems to have closed the door.

Where I stand

This section is my opinion.

I believe the bigger long-term risk is not that children use AI. It's that they use it badly, secretly and without guidance, or that some children get to learn it and others don't. Blanket bans don't make AI go away. They mostly move its use out of the classroom, where teachers could shape it, and into bedrooms and phones, where nobody does.

I also feel that much of the public debate is being driven more by fear than by evidence. That's my view, and I want to be fair about it: many of the people who disagree with me are parents and teachers who care about kids as much as I do, and some of their concerns are backed by real research. I'll get to those below.

We've seen this movie before

When ChatGPT arrived in late 2022, many schools reacted fast. New York City, the largest district in the US, blocked it on school devices and networks in January 2023. By May 2023 it had reversed course. Chancellor David Banks wrote that “the knee-jerk fear and risk overlooked the potential of generative AI to support students and teachers, as well as the reality that our students are participating in and will work in a world where understanding generative AI is crucial” (NBC News). Other big districts, including Los Angeles and Seattle, had also restricted it early on (Business Insider).

Those early bans were understandable. Schools were caught off guard. A UNESCO survey of more than 450 schools and universities in 2023 found that fewer than 10% had any formal guidance on generative AI at all (UNESCO). Three years later, the question is no longer whether AI belongs in education. It's how.

How the world is choosing

Looking beyond one country is useful, because nobody has this figured out, and the range of approaches is wide.

  • UNESCO published the first global guidance on generative AI in education in September 2023. It asks governments to protect student data, train teachers and validate tools, and it sets an age limit of 13 for using AI tools in the classroom (UNESCO).
  • China is building AI into school itself. Beijing required every primary and secondary school to teach at least eight class hours of AI per school year starting in fall 2025 (Gov.cn / Xinhua). At the same time, national guidelines issued in May 2025 bar primary school students from independently using open-ended AI content generators and forbid submitting AI-generated work as one's own (China Daily). In other words: teach it early, but with age limits and rules.
  • South Korea went big and then pulled back. AI-powered digital textbooks rolled out in some grades in early 2025, but after pushback from teachers and parents over screen time, literacy and rushed preparation, the National Assembly downgraded them in August 2025 from official textbooks to optional “educational materials” (Korea Herald; Yonhap). It's a useful warning that rushing in can backfire as badly as shutting the door.
  • Estonia launched “AI Leap,” a national program giving upper-secondary students and teachers free access to AI learning tools, with the stated goal that the benefits “reach everyone, not just schools with better resources” (Estonian Ministry of Education; AI Leap). It has been honest about the hard parts: the student app was delayed by data-protection rules, and by May 2026 only about 63% of eligible students had activated accounts, with weekly use among them at 20% (AI Leap; ERR News).
  • England's Department for Education takes a cautious middle path. It says AI can help every child “regardless of their background,” but that evidence on students using it directly is still emerging, so pupil use should come with close supervision and safety filters (GOV.UK).
  • The United States has moved from silence to active encouragement. An April 2025 executive order made it federal policy to promote “AI literacy and proficiency” and set up a White House task force and a Presidential AI Challenge for students (White House). In July 2025 the Department of Education told schools they may use federal grant money for responsible AI uses like high-impact tutoring and college and career advising (K-12 Dive). At the state level, at least 35 states had published AI guidance for schools as of April 2026 (Education Commission of the States).
  • The OECD will add “Media and Artificial Intelligence Literacy” to its PISA 2029 tests, which compare 15-year-olds around the world (OECD). When the world's best-known school ranking starts measuring AI literacy, that tells you where education is heading.

The pattern I see: almost nobody serious is choosing permanent bans. The real debate is about age, guardrails and pace.

The worries are real. Here's what the evidence says

I want to take the other side seriously, because some of it is well supported.

Cheating is common. About six in ten US teens say students at their school use chatbots to cheat at least somewhat often (Pew Research Center).

Unguided use can hurt learning. This is the strongest evidence on the cautious side. In a randomized trial with nearly 1,000 high school math students in Turkey, researchers from the University of Pennsylvania found that students given a plain ChatGPT-style tool did 48% better on practice problems, but 17% worse than students with no AI on a later exam taken without it. Many had used it as a “crutch” to copy answers. But a version with teacher-designed guardrails, which gave hints instead of answers, largely erased that harm (PNAS, Bastani et al., 2025). To me, that's the most important finding in this whole debate: the design and the teaching matter more than the tool's mere presence.

The “brain” studies are early. You may have seen headlines about an MIT study, “Your Brain on ChatGPT,” which found weaker brain connectivity and poorer recall in people who wrote essays with an AI assistant. It's worth reading, but its authors say it's a preliminary preprint, not yet peer reviewed, with 54 participants from one area, and they specifically ask people not to describe it as AI making us “dumb” (MIT Media Lab).

Students and parents are worried too. In RAND's national surveys, 61% of parents agreed that more AI use will harm students' critical thinking (RAND, 2025), and by December 2025, 67% of students said the same (RAND, 2026). Broader public opinion is cautious too: half of Americans are more concerned than excited about AI in daily life. Across 25 countries the median is 34%, and in no country does excitement come out ahead (Pew US; Pew global).

Privacy is a real issue. Children's data deserves protection. In the US, COPPA requires verifiable parental consent before online services collect personal information from children under 13 (FTC), and FERPA protects school records (US Department of Education). Mainstream tools also have their own age rules. OpenAI, for example, requires users to be at least 13 and to have a parent's permission if under 18 (OpenAI Terms of Use).

Equity cuts both ways. UNESCO warns that today's models mostly reflect the data and values of the Global North (UNESCO). And access is already uneven: US teens in higher-income homes are more likely to use ChatGPT (62% vs. 52%) (Pew Research Center).

Why I still think shutting the door is the bigger risk

This section is my opinion, drawing on the evidence above.

Look at what the research actually shows. The harms come mostly from unguided use. And unguided use is exactly what bans produce, because kids are using these tools anyway. About two-thirds of US teens already use AI chatbots, and three in ten use them daily (Pew Research Center). Yet more than 80% of students in RAND's survey said no teacher had explicitly taught them how to use AI for schoolwork (RAND).

Meanwhile, employers surveyed by the World Economic Forum rank AI and big data as the fastest-growing skill through 2030, and expect 39% of workers' core skills to change in that time (WEF Future of Jobs 2025). The same report stresses that creative thinking, resilience and curiosity are rising too, which is why I want both: kids who can think for themselves and know how to use these tools well.

The equity point matters most to me. When a school bans AI, the kids with tech-savvy parents, tutors and home devices will still learn it. The ones who rely on school for that exposure won't. Estonia's program was built on exactly that concern. A ban can quietly widen the gap it claims to protect against.

What parents can actually do

If your school's approach feels like a closed door, here are practical steps. They work in most places, though the details depend on where you live.

  1. Read the actual policy first. Find your district's AI policy, acceptable-use policy and student handbook. Many “bans” turn out to be narrower than rumored, such as limits on tests or on certain grades. RAND found that in 2025 only 45% of US principals said their school or district had AI policies or guidance, so there may be nothing written yet (RAND).
  2. Start with the teacher, then the principal. Ask, don't accuse: “How do you want students to use AI, and where is it off limits?” Many teachers are already using AI themselves. Six in ten US teachers used it for their work in 2024-25 (Gallup).
  3. Ask for a written opt-in or permission process. If the school worries about age limits or privacy, a signed parental-consent form and an approved, school-managed tool can address both. Ask which tools are vetted and what data they collect. Under FERPA, US parents can review their child's education records (US Department of Education).
  4. Bring official guidance with you. Point to UNESCO's guidance, your state or national education department's AI guidance, and in the US the Department of Education's July 2025 letter. You're not asking for something radical. You're asking the school to follow guidance its own government has published.
  5. Go to the school board. Use public comment time, keep it short and specific, and propose a solution, not just a complaint. Ask whether the district has, or will form, an AI policy committee or task force, and volunteer to serve on it. Many US states have formed task forces of their own (Education Commission of the States).
  6. Teach AI literacy at home. Use AI with your kids: check its answers together, ask it to explain instead of answer, and talk about when it's wrong or biased. Respect each tool's age rules, and choose kid-appropriate tools for younger children. Model the habit you want, which is AI as a tutor, not a ghostwriter. The Turkey study suggests that “hints, not answers” is exactly the habit that protects learning.
  7. Use free, reputable programs. Code.org's “How AI Works” is a free series of lessons for grades 6-12. Day of AI offers free AI-literacy lessons. AI4K12, a joint initiative of the AAAI and the Computer Science Teachers Association, publishes “Five Big Ideas in AI” and a curated list of resources. Libraries, clubs and coding camps often offer more.
  8. Know your wider options. Depending on where you live, there may be transfers, charter schools, online programs, dual enrollment or homeschooling. They vary widely by country and state, so check local rules, and treat leaving as a last resort, not a first move.

The bigger picture

Every generation faces a technology that scares the one before it. The question for humanity isn't whether our children will live with AI. They will. The question is whether they'll grow up understanding it well enough to question it, use it wisely and shape what it becomes, or whether that understanding will belong only to a lucky few.

The evidence so far points to a middle road that is neither fear nor hype: age-appropriate access, strong guardrails, honest teaching about the limits, and lots of human thinking. That's what I want for my kids. I suspect it's what most worried parents want too. We just have to stop arguing about whether the door should be open, and start talking about how to walk through it together.

Sources

#AIinEducation #Parenting #AILiteracy #FutureOfLearning #Education