Teens Wrote a National AI-in-Schools Policy. It Doesn't Ban Detectors — It Makes Them Fairer.

Quick answer: Ninety-eight high school students from every state drafted the STUDENTS FIRST Act of 2026 in late July, a model AI policy for K-12 schools that passed 82-16 at a Boston event hosted by MIT RAISE's Day of AI, AASA, and the Edward M. Kennedy Institute. The act does not ban AI detectors. It requires them alongside human review: teachers must run a standardized detector on graded written work, and any flagged student gets checked by two school officials, not the detector alone. It isn't law yet, but AASA is sending it out to its network of more than 10,000 school leaders.

After many posts on detection scores, the blog keeps one idea: a number sparks a discussion, it never settles the final judgment. Teenagers have written exactly that concept into a model law, resting it in legislation.

What the STUDENTS FIRST Act actually says about AI detectors

By Day of AI the bill is presented in full. Its purpose is to lay out the exact methods for detection, not to argue the necessity of those methods. Of all the clauses, two stand out as the most consequential.

For students, Section 2's first provision reads: "Students should have the right to appeal any convictions of wrongful AI use. They may be subject to technologies such as AI detectors... Students who are suspected of using AI should be guaranteed a thorough examination of their work by two school officials, as well as an AI checker software." NPR's reporting on the event summarized it the same way: any student flagged gets reviewed "not just by AI checker software, which can be unreliable, but also two school officials."

For teachers, Section 4 goes further than most district policies do today: "Teachers must use one standardized AI detector in conjunction with their school district for written graded assessments... If AI detects any misconduct, teachers must personally investigate the issue before reporting the student to admin for improper use." Use the detector, sure. But never let it make the call by itself.

Put those two provisions together and the bill lands somewhere specific. Not "detectors are junk, drop them." Not "trust the score" either. It's the same workflow our own guide to what a detection score means has described all along: run the check, treat the result as one input, then have a human look at the actual work before anyone acts on it.

Why 98 teenagers landed here when adults haven't

Context matters. There is no national policy for AI in schools, so states and districts have improvised on their own, according to NPR. An EdWeek Research Center survey of 693 teachers, principals, and district leaders last fall, reported by Education Week, found most said their school or district had no generative-AI policy, or didn't know if one existed. That vacuum is what the Students First Act is trying to fill.

The students who helped write it weren't naive about what a detector can or can't prove. As Jeff Riley, Day of AI's executive director, put it: the adults haven't reached consensus, so now it falls to the kids to lead. Bena Mejia Romero of Connecticut told her fellow senators before the final vote that the bill was about who gets a seat at the table, and today, students have one. The loudest argument across the three-day event was over when AI belongs in a classroom at all. The group landed here: no AI-assisted writing through eighth grade, brainstorming or editing after that only with a teacher's sign-off, zero AI on any graded test, any grade. Detection enforces those lines, and that's exactly why the bill kept it, just with rules on how it gets used against any one student.

This year Yale, Vanderbilt, and NYU announced that they would no longer use an AI-detector score alone to prove cheating. They made the change after several false-positive cases had caused genuine concern, a shift we've covered in detail. K-12's newest model policy skips the detector-only step entirely. From day one the policy requires detector-plus-human-review. The result is a system designed to keep the same problems from hitting middle and high schools at scale.

What this means if you're a teacher, student, or parent right now

The Students First Act carries no binding effect for districts. Unlike a regulator, AASA is simply a professional association. It issues a template to its members. The template has been received by over ten thousand school leaders across the United States. Most districts confess they lack a real AI policy. The template is copied, whether in full or in fragments, more rapidly than most expect.

A few things actually change once your school adopts something like it, or already runs a similar two-step review.

• A flagged score is the start of a conversation, not a finding. The bill guarantees a human look before any consequence. If you're a student who gets flagged, ask what the review process is and use it, that's the point of writing it into policy.
• Keep your work. Version history, drafts, and notes are the easiest way to make a human review go faster and land correctly, whether or not your school has a formal appeal process yet.
• Detectors report a score, not a certainty. Our own demo shows this directly: paste a passage and you get a probability and a likely-source read, framed as a signal to investigate, the same posture the bill's own teacher provision takes ("personally investigate the issue" before reporting).
• Teachers gain a template for what "responsible use" looks like on paper. A standardized detector, paired with a documented human-investigation step before escalation, is a workable policy shape districts can adapt without waiting on state law.

FAQ

Is the STUDENTS FIRST Act an actual law? The so-called STUDENTS FIRST Act is really just a model policy framework, never a law. It was passed by a mock Student Senate as item 82-16. The framework itself is not binding legislation. AASA is sending it to the schools and districts in its network. Districts can decide to adopt it, tweak it, or leave it behind.

Does the bill ban AI detectors in schools? The legislation does not eliminate AI detectors in schools. It obliges the use of AI detectors for any graded written assignment. Before a student faces discipline, a human must review the detector's findings. This requirement exceeds the standards adopted by many districts. The measure does not represent a rollback of detection practices.

How is this different from colleges dropping AI detectors this year? Colleges got here the hard way. Yale and NYU stopped treating a detector score as final proof only after real students got wrongly accused. Nobody had to get hurt first here, the Students First Act put a human check into the bill from day one.

What should a student do if they're flagged under a policy like this? Ask for the review the policy guarantees. Gather anything that shows your process, drafts, version history, notes, before the meeting. Want to see what a detector actually reports before your case gets reviewed? Run a passage through our free demo, it's a probability, not a verdict.

If you get a detection score from the tool, it'll let you know how much trust to put in it. That score is only a guide, not a verdict. Try the free demo: it only gives you one round of input. Think of the teacher's job, you still need their own eyes on the work before deciding anything.