Adults Have Struggled to Set AI Rules for Schools. When 98 Teenagers Were Asked to Try, They Finished in a Weekend. Here Is What That Tells You About Your Kid.

My oldest daughter is thirteen and has argued against household rules since she could form complete sentences. Last year she made a spreadsheet to challenge our screen time policy. She had research. She had citations. She was a pain about it, and I did not say that to her, because honestly the argument was pretty good. I thought about her this week when I read that 98 teenagers from every state in the country had done something similar to the adults running American schools, except they finished in a weekend and they won.

In late July 2026, 98 high school students representing all 50 states gathered at the Edward M. Kennedy Institute in Boston and passed a comprehensive AI policy for K-12 schools, 82 votes to 16.

The event was organized by AASA, the School Superintendents Association, and Day of AI, a nonprofit launched from MIT’s AI literacy initiative.^[1] The students were given a replica of the U.S. Senate chamber and a weekend. What they produced, the Students First Act, is more specific and more thoughtful than most of the AI guidance that has come out of any adult body in the two years since ChatGPT became a household name.

The bill has real provisions. Students cannot use AI to write. Editing, brainstorming, and studying with AI are permitted after 8th grade. AI is banned on all graded tests. Any student who cites AI in their work must be able to prove understanding through discussion, a handwritten test, or an oral defense.^[1] Schools must teach AI literacy when devices are introduced, covering misinformation, plagiarism, bias, privacy, and environmental impact. Teachers must post written AI policies each semester. If an AI detector flags a student’s work, the teacher must personally investigate before reporting anyone to administration.

The School Superintendents Association plans to share the Students First Act with its 10,000 member school leaders, the majority of whom are writing their own policies right now without any federal guidance.^[1] There is no national AI policy for K-12 schools. The 98 students who showed up for a weekend produced more clarity on the subject than years of adult deliberation have managed to deliver.

Your kid is starting a school year right now with almost no consistent rules for how to use AI, which means the habits that form in the next few weeks will be set by whatever is easy and available.

Only about one in three students nationally says their school has a schoolwide rule for using AI.^[2] The rest are navigating it teacher by teacher, class by class, assignment by assignment, with no consistent standard. That inconsistency does not teach restraint. It teaches kids to read the room and figure out what they can get away with in each context, which is a different kind of learning entirely, and not a useful one.

One of the students who spent the weekend in Boston said something that stayed with me: “We missed the mark with social media and phones. That’s been a huge talking point, is how we failed to implement those things into education.”^[1] She was not blaming adults. She was identifying a pattern and deciding, at seventeen, not to repeat it. That kind of self-awareness is exactly what is missing from the average school conversation about AI, which tends to treat the tool as a problem to manage rather than a technology a kid needs a principled relationship with.

The proof-of-understanding provision in the Students First Act is the one I keep coming back to. The requirement that a student who cites AI must be able to defend their understanding through an oral conversation or a handwritten test is not about catching cheaters. It is about keeping the connection between using AI and actually learning something. The students understood that gap before most adults had named it. And they wrote it into law.

The skill is Self-Governance: knowing your own principles for when to use AI and when not to, before you are in the moment of deciding.

There is a real difference between following a rule someone else made and following a principle you thought through yourself. Both can produce the same behavior in the short term. But the principle holds in situations the rule did not anticipate, and the rule breaks down the moment it is not enforced. The 98 students who drafted and voted on the Students First Act were practicing Self-Governance: they were not waiting to be told what was allowed. They were thinking through what they actually believed before the moment of needing to decide.

For a 9-year-old or a 13-year-old, Self-Governance around AI starts with one question: when is AI actually helping you, and when is it doing something you should be doing yourself? That question does not have a fixed answer. It changes by subject, by task, by where a kid is in their learning. But the habit of asking it, before opening AI and after, is what separates a kid who uses AI deliberately from one who uses it by default. The deliberate version is the one the Students First Act was trying to protect.

A student from Hawaii said during the weekend: “This is for our future, not just for education but for the workforce we will join, the economy that we will build and the communities we will lead.”^[1] That framing is worth holding onto. The AI habits a kid builds in a school year are not just about grades. They are about the kind of thinker they are becoming, and whether that thinker has a clear sense of where their own judgment ends and the AI’s begins.

A child building Self-Governance can tell you why they chose to use AI for something, not just that they did.

One sign is that they have a reason. If your kid used AI for a homework problem and you ask why, they should be able to say something more specific than “because it was faster.” Maybe they used it to check their work after trying themselves. Maybe they used it to understand a concept their textbook explained badly. Maybe they used it to brainstorm and then chose not to use any of what it gave them. Any of those is a reason. A kid who cannot say why is probably using AI by reflex, not by choice, and that is a different thing.

A less obvious sign is pushback. My middle daughter is nine and deeply attuned to what feels right. She pushed back once when I suggested she use AI to help with a creative writing assignment. She said she wanted it to be hers. That instinct, the sense that some work belongs to you and some does not, is an early form of Self-Governance. It was not a rule she was given. It was a line she felt. Kids who come into a school year with that instinct already intact are the ones who will use AI as a tool rather than a substitute, and that distinction is exactly what the 98 students in Boston were trying to write into policy.

Ask your kid: if you could write the rules for how we use AI in this family, what would they be?

Have this conversation before school starts, not after. The habits kids form in the first weeks of a new school year tend to persist. If you wait until a problem comes up, you are already reacting. The 98 students who spent a weekend in Boston were not reacting to an incident. They were setting principles before the chaos of daily decision-making made it harder to think clearly. A 20-minute version of that conversation at the kitchen table before the first day of school is worth more than a longer one you have in response to a problem in November.

Let your kid draft the family rules. Write the AI policy together, but let them lead. Ask what they think AI should and should not be used for at home. Listen to what they say. Push back on things that seem wrong. Agree on things that seem right. The rules you build together will be followed more readily than the ones you hand down, because your kid had a hand in making them. That is the whole lesson from what happened in Boston: when students are included in the policy conversation, they bring better thinking than most adults expect.

Add the proof-of-understanding step. One of the most practical provisions in the Students First Act was requiring students who cite AI to prove understanding through conversation or a handwritten test. You can do a home version: any time your kid uses AI for schoolwork, ask them one question about what they learned. Not as a quiz. As a check-in. If they can answer in their own words, the understanding formed. If they cannot, you know what the next conversation is.

What stayed with me from the Students First Act is not the policy itself. It is the student who said they missed the mark with social media and phones, and decided not to repeat the pattern. She was seventeen. She was sitting in a replica Senate chamber in Boston, and she understood something that most adult AI policy conversations have still not named: the problem with technology is not the technology. It is the absence of a principled relationship with it, formed before the habits set.

My oldest will have opinions about whatever family AI rules we agree on. She will want exceptions. She will argue them well. That is fine. The arguing is part of it. What I want for her is to come out of that conversation with a set of principles she helped write, because those are the principles she will actually enforce on herself. That is what Self-Governance is. And it turns out that 98 teenagers just demonstrated exactly how to build it.