If Your District Banned AI: A Home Curriculum for Grades 3–8

On September 2, 2026, New York City adopted a one-year moratorium on student-facing generative AI for grades 2-K through 8, affecting nearly 600,000 students. Katy ISD in Texas banned generative AI chat tools through sixth grade starting in the 2026-27 year. If your district did something similar, your kid still needs AI literacy, and the ban means you’re the one teaching it. This is an AI literacy home curriculum for grades 3–8 built on real state standards, structured across twelve weeks, where the first six need no account and no chatbot at all.

Key Takeaways

• NYC’s moratorium covers grades 2-K through 8 and includes companion chatbots at all grade levels; high schoolers get five limited pilots and two 45-minute critical-thinking modules a year.
• Idaho’s draft K-12 AI standards, developed by a 27-member committee in July 2026, give you a free grade-band scaffold: K-2 basics, 3-5 daily use and privacy, 6-8 how systems are trained and how bias works.
• Common Sense Media (August 2026) found “only 27% of teens say a teacher has ever talked with them about how to use AI safely.”
• Weeks 1–6 of this curriculum need no AI account. They teach prediction, training data, bias, and verification using paper, family photos, and a deck of cards.
• Total time: about 30 minutes a week. The goal is a kid who can explain what the machine is doing, not a kid who can prompt well.

Why a ban makes home teaching necessary, not optional

A district moratorium on student-facing generative AI removes access at school. It doesn’t remove access, full stop. Common Sense Media’s August 18, 2026 survey of more than 1,000 U.S. teens found “44% have had an AI tool blocked on a school network or device, and 59% of them just switched to a personal device.” That’s the mechanism: the ban relocates the use, unsupervised, to the phone.

NYC’s own policy acknowledges the gap. Per the Mayor’s Office announcement, the moratorium is paired with “twice-yearly AI critical thinking modules” for high schoolers, two 45-minute sessions, plus five pilots capped at roughly 50,000 students. That’s a reasonable start for high school. For a fifth grader, it’s nothing.

And bans aren’t uniform. Katy ISD’s framework, presented by Assistant Superintendent Dr. Sanée Bell in July 2026, bans generative AI chat for K–6 while allowing district-approved curriculum-embedded tools like Amira Learning under direct teacher supervision, then gives seventh graders teacher-directed access only, with broader access in grades 8–12. DC’s OSSE went a different way entirely with a non-mandatory September 1, 2026 model policy governing staff use, using a stoplight system that bars AI from student discipline, teacher evaluations, and IEP eligibility decisions. Illinois issued 409 pages of voluntary guidance in July 2026 that, per Capitol News Illinois, “does not lay out rules or mandates about how schools and districts should use AI.”

Translation: your district’s rules tell you what school will and won’t cover. They don’t tell you what your kid needs. Idaho’s draft standards do that, and they’re public.

The twelve-week curriculum

I built this on the grade-band progression from Idaho’s draft K-12 AI standards, developed by a 27-member committee of educators and industry experts in July 2026 and approved as a framework by the State Board on August 12, 2026 (pending legislative approval, with implementation no earlier than 2027-28). The draft gives grades 3–5 “how people use AI in their daily life, how to decide when to use AI and to protect their personal information,” and grades 6–8 “how AI systems are trained, how bias can affect their outputs, and how to evaluate AI-generated content for accuracy and reliability.”

That’s the spine. Here’s the week-by-week.

Week	Lesson	What you do	Account needed?
1	Prediction, not thinking	Play “finish my sentence” 20 times. Track how often they guess your next word. That’s next-token prediction	No
2	Training data	Look at 20 family photos. Ask: if a machine only saw these, what would it think a “family dinner” looks like?	No
3	Bias from data	Repeat week 2 with photos from one relative’s house only. Name what’s missing	No
4	Confidence vs. correctness	You confidently state three facts, one false. They must identify which. Discuss why confidence isn’t evidence	No
5	Verification habit	Pick any claim from a YouTube video. Find two independent sources. Time it	No
6	Personal information	List what a chatbot could learn about your family from 10 innocent questions. Decide what’s off-limits	No
7	First supervised session	You type, they watch. Ask it something you know well. Catch an error together	Yes, yours
8	Prompting as specification	Same question, three ways: vague, specific, specific with constraints. Compare outputs	Yes, yours
9	Hallucination hunt	Ask for five sources on a topic. Check every one. Log which exist	Yes, yours
10	Study mode vs. answer mode	Ask for the answer, then ask it to tutor without giving the answer. Compare which taught more	Yes, yours
11	Disclosure and citation	Write one paragraph with AI help, then annotate exactly which parts came from where	Yes, yours
12	Their own rules	They draft the household AI rules. You negotiate. Post it	Yes, yours

The account structure matters. Weeks 1 through 6 are deliberately account-free, because most chatbot terms of service set a minimum age of 13, and because the concepts are more durable when taught without the tool in front of you. Weeks 7 through 12 use your account with you present. A grades-3–8 child does not need their own.

How to Teach Your Kid About AI Without a Chatbot

Ages 5–8: The guessing machine

Sit with a picture book you’ve read fifty times. Read a sentence and stop before the last word. Let them fill it in. They’ll be right most of the time. Ask how they knew. “Because that’s how it goes” is exactly the right answer, and it’s exactly what a language model does. Then read a brand-new book and stop mid-sentence. They’ll guess wrong more often. That’s a model outside its training data, explained to a six-year-old in four minutes.

Ages 9–12: The card-deck bias experiment

Take a deck of cards and remove all the hearts. Hand your child ten cards from the remaining deck and ask them to guess what a “typical card” looks like. They’ll never say hearts. Now tell them hearts exist. The reaction is the lesson: a system’s picture of the world is made of whatever it saw, and what it never saw doesn’t feel missing. Then apply it: what groups of people, accents, or places might a chatbot have seen very little of?

Ages 13+: The source audit

Have your teen ask an AI tool for five sources on a topic they care about, then check all five independently. Log each as: real and says what the AI claimed, real but misrepresented, or doesn’t exist. Then have them repeat the exercise on a topic where they’re already an expert (a game, a sport, a band). The error rate on familiar territory is usually the moment it clicks. Weber-Wulff and colleagues (2023, International Journal for Educational Integrity) tested 14 AI-detection tools and concluded they are “neither accurate nor reliable,” which is a useful parallel: the tools around AI are as fallible as the AI.

The question to ask: “How would you know if it made this up?”

Running the curriculum without it collapsing

Attach it to something that already happens weekly

Thirty minutes, same slot, tied to an existing anchor: Sunday breakfast, the drive to practice, after-dinner cleanup. Lally, van Jaarsveld, Potts, and Wardle (2010, European Journal of Social Psychology) found habits reached maximum automaticity at a median of 66 days, with a range from 18 to 254. Twelve weeks is 84 days. Attaching the new behavior to an existing cue is the mechanism that makes it stick.

Let them catch you being wrong

The most valuable moment in the whole twelve weeks is the one where your kid catches an error you didn’t plan. Don’t defend. Say “good catch” and move on. The skill you’re transferring is skepticism toward confident sources, and you are a confident source.

Keep your account, not theirs

Through eighth grade, use your own account with you in the room. This isn’t about surveillance; it’s about the fact that OpenAI’s teen experience and its parental linking apply to accounts for 13- to 17-year-olds, and a fifth grader is below that floor entirely. When they turn 13, that’s the conversation about setting up a teen account properly.

Write the rules with them in week 12

The final week isn’t a quiz; it’s a negotiation. Have your child draft the household AI rules, then edit together. Rules a kid wrote get followed at a rate that rules imposed on them don’t. This is also where you discover what they actually believe about the tool.

What not to do

Don’t teach prompting as the core skill. Prompt techniques change every few months; the mechanism doesn’t. A kid who understands that the model predicts plausible text from patterns will adapt to whatever interface ships next. A kid who memorized prompt formulas will not.

What to Watch For Over the Next 3 Months

• Week 4: Can your child explain “it’s guessing what comes next” in their own words? That’s the load-bearing concept for everything after.
• Month 2 red flags: The weekly slot has become “let’s just ask the AI.” Also watch for the opposite: a kid who’s become dismissive and won’t engage with weeks 7–12.
• Month 3 self-check: Hand them an AI-written paragraph on a topic they know. Can they find the error without being told there is one?

Frequently Asked Questions

Which districts have actually banned student AI use?

New York City adopted a one-year moratorium on student-facing generative AI for grades 2-K through 8 on September 2, 2026, affecting nearly 600,000 students. Katy ISD in Texas bans generative AI chat tools through sixth grade for 2026-27, with teacher-directed use in seventh and permission-based access in grades 8–12.

Does my kid need a chatbot account to learn AI literacy?

Not for the first half of this curriculum. Weeks 1 through 6 teach prediction, training data, bias, confidence versus correctness, verification, and privacy using paper and conversation. Weeks 7 through 12 use a parent’s account with the parent present.

What should a fifth grader actually know about AI?

Idaho’s draft standards for grades 3–5 target how people use AI in daily life, how to decide when to use it, and how to protect personal information. Grades 6–8 add how AI systems are trained, how bias affects outputs, and how to evaluate AI content for accuracy.

How long does this take each week?

About 30 minutes. The point isn’t volume; it’s repetition attached to a consistent cue. Habit research suggests a median of about 66 days to automaticity, with wide individual variation, so twelve weeks is a reasonable first commitment.

Will this put my kid behind if their school stays banned?

It puts them ahead of the typical case. Common Sense Media found only 27% of teens report a teacher ever discussing safe AI use. A child who has done twelve structured weeks at home is in a small minority regardless of district policy.

Is a ban a bad policy?

Reasonable people disagree, and the evidence is genuinely mixed. PISA 2025 found moderate AI users outscored both non-users and near-daily users, which argues against both extremes. Bans do remove school-day access, and surveys show many students shift to personal devices, which is the strongest practical argument against them.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. NYC Mayor’s Office. (Sept 2, 2026). “Mayor Mamdani and Chancellor Samuels put students first with nation’s broadest generative AI moratorium in schools.” https://www.nyc.gov/mayors-office/news/2026/09/mayor-mamdani-and-chancellor-samuels-put-students-first-with-nat
2. Idaho EdNews. (July 13, 2026). “Idaho’s AI draft standards aim to teach students when — and how — to use the technology.” https://www.idahoednews.org/news/idahos-ai-draft-standards-aim-to-teach-students-when-and-how-to-use-the-technology/
3. Idaho EdNews. (Aug 12, 2026). “State Board roundup: AI standards approved for K-12 schools.” https://www.idahoednews.org/news/state-board-roundup-8-12-26-board-approves-ai-standards-for-k-12-schools/
4. Common Sense Media. (Aug 18, 2026). “Teens in the AI Era: Schoolwork and Skills That Matter.” https://www.commonsensemedia.org/research/teens-in-the-ai-era-schoolwork-and-skills-that-matter
5. Covering Katy News. (Aug 11, 2026). “Katy ISD’s new AI rules: What parents and students need to know.” https://coveringkaty.com/education/katy-isd-s-new-ai-rules-what-parents-and-students-need-to-kn/
6. OSSE (District of Columbia). (Sept 1, 2026). “OSSE releases AI model policy to guide responsible staff use in schools.” https://osse.dc.gov/release/osse-releases-ai-model-policy-guide-responsible-staff-use-schools
7. Capitol News Illinois. (July 15, 2026). “Illinois State Board of Education issues AI guidance, written with help from AI.” https://capitolnewsillinois.com/news/illinois-state-board-of-education-issues-ai-guidance-written-with-help-from-ai/
8. Weber-Wulff, D., Anohina-Naumeca, A., Bjelobaba, S., et al. (2023). “Testing of detection tools for AI-generated text.” International Journal for Educational Integrity, 19(26). https://doi.org/10.1007/s40979-023-00146-z
9. Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). “How are habits formed: Modelling habit formation in the real world.” European Journal of Social Psychology, 40(6). https://onlinelibrary.wiley.com/doi/10.1002/ejsp.674

Related reading: what AI literacy actually means for a 10-year-old, NYC’s Pre-K–8 AI ban explained for families, and Idaho’s grade-by-grade AI learning standards.