AI & Parenting

A School AI Pause Isn’t a Home Playbook

What the news, two surveys and three pieces of research do and don't tell parents and teachers about homework and AI, plus one coaching idea of mine.Some school districts are…

What the news, two surveys and three pieces of research do and don’t tell parents and teachers about homework and AI, plus one coaching idea of mine.

Some school districts are pausing student-facing generative AI. When there’s no policy yet, that’s an understandable response. But a pause covers school tools, and homework doesn’t only happen at school. So what can a parent or teacher do about AI use outside the pause?

What’s being reported

Separate reports, not a national picture.

  • The NYC Mayor’s Office announced on September 2 a one-year moratorium, effective in the 2026-27 school year, on student-facing generative AI for children in 2-K through 8th grade, which it says affects nearly 600,000 students. Companion chatbots are prohibited in all grades. NYC Schools Chancellor Kamar Samuels is quoted in the release: “…over the next year, we will lead with evidence to make sure technology serves learning…”
  • Los Angeles Unified restricts all students’ access to generative AI on district-issued devices for the 2026-27 school year, according to a district statement reported by NBC Los Angeles and K-12 Dive, and by the Los Angeles Times via EdWeek. Administrators put it in place as an administrative practice, not a board vote, and it came to light at a September 2 board committee meeting. The district has no policy yet on personal devices.
  • Beaverton Beat reported on October 15 that around 50 parents, students and educators with Schools Beyond Screens rallied on October 13 for a pause on generative AI in classrooms until the Beaverton School District adopts a formal policy. The article says the district has no board-approved, district-wide AI policy, and the group’s petition (over 1,000 signatures) asks for parent training, among other things.
  • The Acorn reported on October 16 that the Las Virgenes Unified board approved a ban on large language models and generative AI tools in TK through fifth grade, with middle and high school use under teacher guidance.

What kids say about AI and schoolwork

Common Sense Media published two self-reported surveys, not tests of learning.

  • Its Census (1,204 U.S. kids ages 9 to 17, fielded March 18 to 26, 2026, margin of error about 3.8 points) found 86% use or interact with AI, including 89% of 13- to 15-year-olds. Its definition of AI is broad and includes things like AI summaries in search results. Among kids who use AI, 85% said they use it for help with schoolwork or homework.
  • A separate nationally representative survey, “Teens in the AI Era: Schoolwork and Skills That Matter” (press release August 18, 2026), was run by NORC at the University of Chicago for Common Sense’s Youth AI Safety Institute: 1,017 teens ages 13 to 17, interviewed April 30 to May 14, 2026, margin of error +/-4.3 percentage points. It found 70% of teens use AI for schoolwork. Among teens who use AI for schoolwork, 63% say they get an answer from AI. The report’s headline words this as “Among teens who use AI, 63% say they get an answer to their schoolwork from AI,” and the chart’s base is the 665 teens who use AI for schoolwork (a subgroup, so the margin is wider). It also says more than three quarters (77%) of teens who use AI for schoolwork say they use it for help without getting direct answers. And 73% of teens say teachers have not talked with them about what AI is and how it works.
  • The report also finds that 44% of teens who use AI for schoolwork had an AI tool “blocked or banned” on a school network or school-issued device, and 59% of those teens used a personal device to access AI instead.

That last finding is self-reported: it is what teens told the survey, not evidence that a pause causes home use. The survey was fielded in spring 2026, before the fall 2026 NYC and LA moratoriums, and it asked about tools blocked or banned, which may not be moratoriums like the ones above.

What the research says about design

  • Bastani and colleagues (PNAS, 2025). A randomized experiment with nearly 1,000 high-school math students at one school in Turkey, Fall 2023. During practice, students with a standard chat tutor scored 48% higher than the control group, and students with a tutor built to give teacher-designed hints scored 127% higher. On the later closed-book exam, the standard-chat group scored 17% lower than students who never had AI, while the hint-based group was statistically indistinguishable from the control group. The authors describe students using the standard tool as a “crutch,” often asking for and copying solutions.
  • University of Minnesota expert consensus report (October 2026). Thirty experts took part in a structured consensus process on 65 learning processes. Their first recommendation is “Epistemic Agency First”: learners should engage in independent thinking, drafting, or problem-solving first before consulting AI.
  • AERO (Australia, October 2026). In provisional guidance, AERO says AI should scaffold learning “rather than replace the cognitive work students need to do themselves,” and gives “a sequence of probing questions that prompt students’ thinking rather than providing answers directly” as an example.

How much to trust it

  • The news items are separate reports of what individual districts did, not a national count. The NYC details come from the city’s own release; the LA details come from press accounts of the district’s statement.
  • “Uses AI” is broad in the surveys, and the 85% is a share of AI users, not of all kids.
  • Bastani is one school, one subject, short-term. The authors note habits may have changed since. Their hint tutor also had teacher-written solutions and hints built in. A parent typing “just give hints” into a chatbot is not the same thing, and I can’t say my routine protects the way their tutor did.
  • The Minnesota report is expert consensus, not an experiment, and says strong evidence is still lacking. AERO says there is very little evidence on AI in school education.

My own interpretation: a home coaching layer

This part is mine. None of the sources tested it, and none of them endorse Auxesis.

My two questions are whether AI augments a child’s thinking or replaces it, and whether the work deepens understanding or shortcuts it. Read that way, the research points to one thing: the order and shape of the help matter. Here is a routine I’d try, for kids old enough to use these tools at all. Check each tool’s own age rules first: many popular chatbots require users to be 13 or older (some 18), and some need a parent’s permission under 18. Keep your child’s name, school, photos and voice recordings out of any prompt. Minnesota calls elementary grades the most vulnerable, and AERO urges greater caution before secondary school, so for younger children I’d lean toward adult-led use or none.

  1. Think first. The student writes their own attempt before any AI, even a messy one.
  2. Hints, not answers. If AI is used, ask for a hint or a question, never the answer or the finished text. The student does the step.
  3. Ask the AI to question you. “Ask me three questions about my attempt,” not “fix it.”
  4. Closed-book check the next day. One problem or one explanation out loud, no tool in reach. Bastani’s students didn’t notice their own loss, so I wouldn’t rely on how it feels.
  5. Say what you used. Tell a parent or teacher which tool, and where you got stuck.

As a child shows they can do steps 1 and 4 alone, I’d fade the support. And I’d treat it as a conversation, not a ban.

Try this

Which of those five steps would be hardest to hold at your house, and why? Tell me which, and what you saw. (Please don’t share names or identifying details about your kids.)

Continue this in the community: https://auxesislab.com/community/

Sources

Leave a Reply

Your email address will not be published. Required fields are marked *