27 States Are Writing AI Rules for Your Kid's School. I Run One. Here's What They're All Missing.

27 States Are Writing AI Rules for Your Kid's School. I Run One. Here's What They're All Missing.
 
  Twenty-seven states have active AI-in-education bills in 2026. Five are already law. Ohio's House Bill 96 required every public, community, and STEM school district in the state, more than 600 of them, to adopt a formal board-approved AI policy by July 1, 2026. California released a statewide model AI policy for districts on June 25. Virginia's law took effect July 1. And on August 5, the Florida State Board of Education takes up a rule that would require public and charter schools to address artificial intelligence in their internet safety policies.
 
  M.A. Aponte is the Executive Director of a charter school in Gainesville, Florida. That Florida rule governs his building.
 
  So this is not policy analysis from the outside. This is a practitioner explaining what happens when a mandate lands on a real school: what he has to tell teachers, how he answers parents, and what he does when a student turns in an essay a machine wrote.
 
  His concern is simple and uncomfortable. A mandate produces a document. It does not produce a change in practice. And if districts confuse those two things, the next two years will be spent celebrating paperwork while nothing improves for a single student.
 
  WHAT THIS EPISODE COVERS
 
  The full national map of school AI legislation, including Ohio's July 1 deadline, California's model policy, Idaho's AI literacy standards, Oklahoma's parental opt-out provision, Maryland Senate Bill 720 requiring teacher professional development, Connecticut's AI instruction mandate, and South Carolina House Bill 5253, which would establish the strongest guardrails in the country.
 
  What is happening in Florida specifically, including the Department of Education's internet safety directive, the August 5 State Board workshop, and the AI Bill of Rights that passed the Florida Senate 37 to 1 before dying in the House.
 
  The strongest argument against moving this fast. MIT professor Justin Reich has said that writing a guide for AI in 2026 is like writing a guide for aviation in 1905, before there were airlines, airports, or commercial flight. Aponte takes that seriously and explains where he lands anyway.
 
  The four things most school AI policies get wrong:
  1. They regulate tools instead of teaching thinking. The real question is not whether a student used AI. It is whether AI helped them think or thought for them.
  2. They assume detection works. AI detection tools flag authentic student work at rates that make them unusable as the basis for a disciplinary finding.
  3. They do not fund the training. Teachers repeatedly report having to navigate AI entirely on their own.
  4. They confuse having a document with having made a change.
 
RESOURCES
 
  Free AUDIT Framework worksheet, including a parent version with five questions to bring to a school board meeting: maaponte.substack.com
 
  School AI policy consulting and advisory services: link in description.
 
  Sources cited include the FutureEd 2026 State AI in Education Legislative Tracker, Stateline, the California Department of Education, the Florida Department of Education, the OECD Digital Education Outlook 2026, UNESCO, and research from the University of Florida Education Policy Research Center.

FREQUENTLY ASKED QUESTIONS

 Q: Which states require schools to have an AI policy in 2026?

  A: As of mid-2026, 27 states have active AI-in-education bills and five have enacted laws. Ohio House Bill 96 required every public, community, and STEM school district to adopt a board-approved AI policy by July 1, 2026. Idaho requires districts and charter schools to create AI policies and ensures AI cannot replace a human teacher. Oklahoma requires age-appropriate AI tools, teacher review of AI output, and parental opt-out. Maryland Senate Bill 720 requires state-provided teacher professional development on AI. Virginia Senate Bill 394 directs its education department to issue guidance. Connecticut mandated AI instruction. California released a model AI policy for districts on June 25, 2026.

  Q: What should a school AI policy include?

  A: According to the AUDIT framework developed by charter school director M.A. Aponte, an effective school AI policy must address five areas. Assumptions: does it reflect how students actually use AI, including gradual over-reliance rather than only deliberate cheating. Usability: can a teacher apply it in real time without stopping instruction. Detection: does it acknowledge that AI detection software produces false positives, and does it rely on process artifacts and oral defense instead. Instruction: does it teach AI literacy rather than only restricting tools. Training: is there a funded, dated, ongoing professional development plan for teachers.

  Q: Is Florida requiring schools to have an AI policy?

  A: The Florida Department of Education has directed school districts to add artificial intelligence use to their existing internet safety policies. A rule development workshop was held June 22, 2026, and the Florida State Board of Education scheduled a workshop for August 5, 2026 on a proposed rule requiring public and charter schools to include AI in their internet safety guidelines. Separately, Florida's proposed AI Bill of Rights, which would have restricted student access to AI instructional tools before sixth grade, passed the Florida Senate 37 to 1 but died in the House.

  Q: Do AI detection tools work in schools?

  A: AI detection tools produce false positives at rates that make them unreliable as the basis for academic dishonesty findings, flagging authentic student work as AI-generated. Education leaders recommend assessment design strategies instead of detection strategies: requiring drafts and revision history, conducting device-free in-class writing, and asking students to orally defend submitted work. A student who wrote the work can explain their reasoning. A student who outsourced it cannot.

  Q: What questions should parents ask their school about AI?

  A: Five questions. Does our school have an AI policy, and can families read it? Has a classroom teacher reviewed it for whether it is usable during instruction? What happens if a student is wrongly accused by AI detection software? What are we teaching students about AI, not just what are we forbidding? When are teachers trained on this, how often, and who pays for that time? If a school cannot answer the last question, that is the gap that determines everything else.

  Q: Why do most school AI policies fail?

  A: Four reasons. They regulate tools instead of teaching thinking, so they cannot distinguish between a student using AI to learn and a student using AI to avoid learning. They assume detection is possible when detection software is unreliable. They do not fund teacher training, leaving educators to navigate AI on their own. And they confuse producing a compliant document with changing classroom practice. A mandate produces a document. It does not produce a change in practice. 

Transcript

SPEAKER_00 0:02

Every school I have ever worked in has a binder. Sometimes it's a shelf, sometimes it's a shared drive folder that nobody has opened since the day it was created. Inside our policies, dozens of them, written carefully, approved formally, filed properly. And a meaningful number of them have never changed a single thing about what actually happens in a classroom. And that's not cynicism. That's just what happens when a policy is written to satisfy a requirement instead of trying to solve it. The documents exist, the behavior doesn't change, and everyone can point to the binder and say, we addressed it. And I'm telling this because right now, across the United States, that exact process is happening with artificial intelligence at an enormous speed, at an enormous scale. 27 states have an active AI and education bill this year. Five are already in law. Ohio required every public communities and STEM school district in the state to adopt a formal board-approved AI policy by July 1st. That deadline hit less than a month ago from this recording. California released a motto policy for districts for June 25th. Virginia's new law took effect uh July 1st. Nationally, lawmakers introduced more than 1,500 AI-related bills this year. And here in Florida, where I run a charter school, the state Board of Education has a workshop scheduled for August 5th on a rule that would require public schools and charter to address artificial intelligence in their internal safety uh policies. That's about nine days from when this episode gets published. That rule would govern my building. So I'm not covering this story as a commentator. I'm gonna make that clear. I'm covering it as a guy who has to implement it, who has to explain it to teachers, who has to answer parents about it, who has to decide what happens when a 15-year-old turns in an essay that a machine wrote. Here's my concern. And I want to be careful how I say it, because I'm not against these policies. I think the states are right to act. What worries me is this a mandate produces a document. It does not produce a change in practice. And if we confuse the two, we will spend the next two years congratulating ourselves for paperwork while nothing improves a single student. Today I'm going to walk you through what's actually happening, what most of these policies get wrong, and give you a five-question tool, the audit framework, that lets any parent, teacher, or school leader tell the difference between a policy that will work and a policy that will sit in a binder. So don't forget to like and subscribe before I even continue this. Um also I was accused of being AI, so I am not. So I decided to do this episode in video. And uh if you're watching a video, fantastic. Now you know I'm not AI. If you are uh listening in audio, then there's a video version of just me speaking. And you know, it is what it is. But uh without further ado, let's establish the facts before we get to opinion. That's how we do it here in uh thinking to think. Two years ago, uh, most schools had nothing. A you know, a survey of more than 450 schools and universities worldwide found that only about 10% had established any guidelines for AI use. 90% were flying blind while their students used these tools every day. Uh, that has changed fast. Uh, according to the feature ed legislator tracker, 52 bills across 25 states now address artificial intelligence uh in classroom instruction. What students learn about AI, how schools integrate it, and what rules govern it uses. Uh, the PIE or PI network tracked nearly uh 100 state bills uh this year that could directly affect how students use AI in K through 12. And those sit inside a broader wave of more than uh 1,500 AI-related uh bills introduced by lawmakers nationwide. So, in roughly 24 months, we went from almost no schools having a policy to states legally requiring them. That is remarkable turnaround in a sector that usually moves very slowly due to bureaucracy. So credit where credit's due. Um now, these are the states that moved the fastest. All right, whole Ohio. Ohio went furthest fastest. House Bill 96 required every public community uh and STEM school district in the state to adopt a formal board-approved AI for policy by July. And the approach uh is worth understanding because other districts are copying it. Uh Columbus City Schools adopted this earlier. Um, they treat AI as a learning supplement, not a substitute for a student uh effort or teacher judgment. They give teachers full discretion uh to decide whether AI can be used on any given assignment. Unauthorized use is treated as plagiarism, and critically, they bar student and staff from being used to train AI models. Um, Idaho's new law requires local districts and charter schools to create AI policies, uh sets uh state standards for AI literacy, and this is an important part, explicitly ensures that no AI replaces or eliminates a human teacher. Oklahoma requires that AI tools be age appropriate, requires teachers to review anything AI produces before using it in the classroom, and allows parents to opt their children out of AI tools entirely. Maryland went a direction I find genuinely smart. Uh, Senate Bill 720 requires the state to provide professional development for teachers uh on AI, not just rules, training. We'll come back to why that matters more than almost anything else. Uh, Connecticut mandated AI instruction. Virginia directed its Department of Education to create guidelines, will eventually require local school boards to adopt their own California. Uh California released a ready-to-adopt model policy on June 25th, built by working group uh convened under state law, uh covering legal, ethical, and logistical considerations for any school AI policy. And South Carolina has proposed what may be the strongest guardrails uh in the count in the country. House Bill 5253 would require written parental opt-in consent, annual public disclosure of what AI tools a school uses, and how student data is handled, and would prohibit AI from replacing licensed teachers in core instruction or in grading. It would ban automated high-stakes decisions without meaningful human oversight and restrict student profiling. So now we get to Florida. So what's happening in Florida? Now I'm gonna bring it home. Uh, and this is where I currently work. Uh Florida took uh a narrow path. Uh the Florida Department of Education is directing school districts to add artificial intelligence to use their existing internet safety policies. Uh, every student district already has an internet safety policy on file. The DOE is requiring that it amended to include AI. There was a rule development workshop on June 22nd, and the state board of education has scheduled another workshop for August 5th on the proposed rule change, which would cover public and charter schools just weeks before students returned to classroom. Uh, just so you know, this is August 5th, and um in my county, we start the following week. So um yeah, uh there was it was also a bigger swing that didn't land. Uh Florida proposed the bill, um, the right that would have restricted um restricted student access to AI, instructional tools, uh, before sixth grade, uh, with exceptions for supervised use, uh, transl translation support for English learners and disability accommodations. It passed the Florida Senate uh 37 to 1. Um that's nearly unanimous and then died in the House. Uh so Florida's current posture is fold AI into the internet safety policy you already have. That's the real step. Okay. It is also, in my honest assessment, as someone who has to implement it, um it's a modest one, it's not not terrible. Um, an internet safety policy is designed to answer uh a different question than the one AI possesses. Internet safety um asks, what content should students um be protected from? AI asks, what thinking should students be doing themselves? Those are not the same question. And a policy built to answer the first one would not automatically answer the second. One more Florida note, and it's local for me. Researchers at the University of Florida, which is Gainesville, uh analyzed 73 school district codes of conduct from the 24-25 uh school year to see how Florida districts govern student AI use, including how they connect it to academic dishonesty and technology misuse. The research exists, it's public, and I'm encouraging any Florida leader uh listening to go write and read it before they write their own policy. So here's the honest counterpoint. And before I critique any of this, I want to steel man the case for caution. The strongest version of the argument that we should be careful about rushing to write rules at all. Uh, Justin Reich, an associate professor of digital media at MIT, made a point that stuck with with me. He said the uncertainty around AI should make policymakers cautious uh about declaring best practices soon. And his line was that writing a guide uh for AI in 2026 is like writing a guide for aviation in 1905, before there were airlines, airports, or even commercial flight. And that's a serious argument, and I don't dismiss it. If you write rig uh rigid rules right now, you may be locking schools into assumptions about technology that will look completely different in 18 months. That's a real risk of codifying today's misunderstanding into tomorrow's requirement. So here's where I land. And I had both of these at once. The absence of policy is also a policy when there are no guardrails or guidelines, excuse me, the decision doesn't disappear. Uh, it just gets pushed down to the individual teacher in the moment with no support. And more often than not, uh it gets pushed down further than that to a 15-year-old at 11 o'clock at night deciding alone whether to paste an assignment into a chatbot. So, yes, uh, write the policy, just write one that expects to be revised. The mistake isn't acting early, the mistake is acting once and calling it finished. So, the four things most policy, most AI policies get wrong. Uh, and I want to break this down as a practitioner. These are not attacks on people writing these policies. Please make I want to make that abundantly clear. Most of them are working hard on a genuinely difficult uh problem with limited time. These are the gaps I keep finding. So, Wong number one, they regulate tools instead of teaching thinking. Most AI policies are written as tool policies, which apps are approved, what's blocked on the network, uh, what counts as unauthorized use, what consequence is when the student violates it. All of that is necessary. None of it is sufficient because the actual educational um problem is not with the software the student opened. The actual problem is whether the student is developing the ability to think. I talked about this in an earlier episode, and I'll restate it in one sentence. The question is not whether a student used AI, the question is whether AI helped them think or thought for them. Those produce very different students. In the tool policy, you usually can't tell the difference. So here's the here's why the distinction matters so much. The OECD, that's the organization for economic uh co-operation and development, essentially the research body for the world's uh developing nations uh documented what they called a mirage of false mastery. Students using AI that hands them the answer show dramatic short-term uh score improvements. Uh, when the AI was removed, their performance dropped below where they started. Started uh students using AI that gave hints and made them work through the problem themselves retained gains. Retained gains. Same technology, opposite outcomes. The variable wasn't a tool, it was the design of the interaction. A policy that only names approved apps cannot capture the difference, and the difference is the entire ball game. Number two, they assume detection works. Um, a lot of these policies uh treat unauthorized AI use as plagiarism, reasonable on its face, but that framing uh quietly assumes something that isn't true, that we can reliably detect it. AI detection tools flag honest student work uh as an AI generated at a rate that should stop us cold. We are talking about false accusations against students who did their own work at a frequency no school should accept as routine. If you don't believe me, if you have a published book, well uh, and it could be any of the even the uh modern textbooks, and you scan a page and you upload it to an AI detector, there's a high probability it will come back. And this is before AI was available to common people, it would detect that it was written by AI. So if you take a book from the 90s, we'll say Harry Potter, and you scan it, just and put it in the AI detector. That's copyright, be careful. Um just for fun and games for an experiment, there's a high probability, it would say it's a almost a high probability was written by AI. So be mindful of that. Um think about what that means in practice. A teacher runs an essay through a detector, it comes back flagged. Now the teacher has to choose between trusting a piece of software or trusting a kid. And whichever way they choose, they might be wrong. And the student pays for it either way. I've had to sit in that room. It's one of the worst positions you can put an educator in. And a policy that says unauthorize and use equal plagiarism without addressing how detection actually works is handing teachers an impossible job with a straight face. A policy that assumes detection works will produce false accusations, damage trusts, and eventually teachers would just stop enforcing it at all because the cost of being wrong is too high. That's how a rule dies quietly, not by being repealed, but by being ignored. Three, they don't fund the training. This is the one that frustrates me the most, and it's the most fixable. You can write the best AI policy in the country. If you hand it to teachers with no training, no time, no support, you have not implemented anything. You have distributed a document. You can write the best AI policy in the country. If you hand it to teachers with no training, no time, no support, you have not implemented anything. You have just distributed that document. I cannot stress that enough. A Maryland State Senator, uh, Katie uh Hester, and I could Fai Hester, I could be saying her name wrong, and I apologize, says something in a reporting on this that I think is the truest sentence in the entire national conversation. She said what they heard repeatedly was that teachers felt like they had to navigate artificial intelligence entirely on their own. Entirely on their own. That's not a policy failure, that's an implementation failure. And they are not the same thing. I during orientation and some professional development, I train my teachers on how to use AI for themselves and their practices to simplify their work, streamline, and use it effectively. And I plan uh when the teachers come back to the school to elaborate on that training and how to make their how to streamline, give them the skill set that they can use AI as a benefit and give them more time, not only for themselves, because of burnout is real, but also to more individualized instruction. Because every study, time and time again, individualized instruction helps students. So credit to the states that saw this. Uh, Maryland Senate Bill 720 requires the state to offer teachers professional development related to AI. Uh, Virginia Senate Bill 394 directs its state education agency to provide guidelines, including teacher training. Uh those provisions who will well do not do more real good uh than any list of approved applications. Um because here's the operational truth from inside a school a policy is only as good as the adult standing in front of 30 students who has to apply it in real time with no time to consult a binder while also teaching a lesson. If that adult hasn't been trained, the policy is theoretical at best. Number four, they confuse the document with the change. This is the big one, and it's one of the reasons I opened this episode um with the binder. When a state sets a deadline, districts organize around the deadline, they the goal quietly shifts from make our classroom uh better to have a compliant document by July 1st. Those are different projects. They can overlap, they often do not. There's a policy uh consultation, cult consultancy, uh tracking these mandates that made a point uh worth repeating. Template alone does not satisfy the framework requirement in states like Idaho, Ohio, and Georgia. Uh those laws require evidence of operational governance, uh defined roles, vendor protocols, uh staff training, data process uh practices, excuse me, not just a document on file. And I want school leaders specifically to hear that clearly. Because I think I think a lot of districts are going to be surprised. Uh downloading California's model policy and putting your district's name on it, it's a starting point. It's not compliance. And it's entirely and certainly isn't leadership, it's just copy paste. And I'll say the uncomfortable part out loud because I felt the pull myself. When you're a school leader, stretch thin, a compliant document is enormously tempting. It's finishable. Uh, you can check it off. Actual implementation is never finished. Nobody hands you a certificate for it, but it's the only part that reaches a kid. And also, I won't get to that. Uh so here's my contribution. You if you heard from this this long, and it's not an audit, I mean it's not a rant or anything like that, but it um if you're listening, I'm gonna give you something that's deliverable. Here is my contribution to you. I built this for my school, and I'm giving it to you because I think the national conversation uh needs a practice tool and not another think piece. It's five questions. Any parent can ask them at the school board meeting, any teacher can ask them in a staff meeting, any school leader can run them against their own draft. And I call it, and it's an acronym, audit framework assumption, usability, uh, detection, instruction, and training. So audit. All right, here we go. Assumptions. The question what does this policy assume about how students actually use AI? And is that assumption true? Most policies are written imagining a student who deliberately decides to cheat, sits down, opens a chat bot, generates an essay, submits it. That student exists, but that's not the common case. The common case is far blurrier. Uh a student uses AI to understand a confusing paragraph, then to check whether their thesis makes sense, then to rephrase an awkward sentence, then to expand a thin sentence. At no point they did they decide to cheat. There were never a moment where the line got crossed on purpose, it's eroded gradually. A policy written for the deliberate cheater will not help that student. And it won't help the teacher trying to guide them. Check the assumption first, because everything downstream inherits it. So, assumption. What does this policy assume about how students actually use AI? And is that assumption true? U, usability. Here's the question: Can a teacher actually apply this in a live classroom in real time without stopping instruction? This is uh this test, almost every policy fails. Um, and it's one of the that nobody applies before adopting. Read your school's AI policy and then picture second period. 30 students, a teacher who has roughly four seconds to make a judgment call before that moment, uh, passes and the classes move on. If applying the policy requires a teacher to consult a document, weigh three three conditions, excuse me, and check for approved tools uh list, it will not be applied. Not because the teachers don't care, because the classroom doesn't pause. Usable policy fits in the sentence a teacher can hold in their head. Columbus City schools got closer to this than most. Um, teachers have full discretion to decide whether AI may be used on any given assignment that's usable. A teacher can carry that, compare it to a four-page decision tree, and you'll see the difference immediately. So, usability. Can a teacher actually apply this in a live classroom in real time without stopping? D, detection. Here's the question: what happens when the policy is violated? And is detection actually possible? If your policy has a consequence, it needs to, it needs an honest assessment about how the violation got identified. And given what we know about detection tools accuracy, most policies do not have one. Here's my position. And school leaders can disagree with me. Totally fine. Free country, up to you. If you cannot reliably detect a violation, do not build your policy on catching people. Build it on making the work harder to outsource in the first place. That means process the artifacts, uh, drafts, outlines, uh, revision history, in-class writing without devices, oral defense of submitted work. Walk me through your argument. Why did you pick this evidence? What's the strongest objection? A student who voted can answer, a student who outsourced cannot, and no software has to accuse anyone of anything. That's not a detection strategy, that's an assessment design strategy, and it's the only one I found that actually holds. So detection D. What happens when the policy is violated? And is the detection actually possible? I instruction. The question: Does this policy teach thinking or does it only restrict tools? Restriction is a floor, not a ceiling. If everything in the policy is about uh what the students may not do, you have a discipline document, not an education document. The states that understand this went further. Uh, Connecticut mandated AI uh instruction, not just AI rules. Um Idaho set the state uh standards for AI literacy alongside its policy requirement. That's the difference between telling students what's forbidden and teaching them how to think in a world where this technology exists, whether we like it or not. That Pandora box is open, it is not closing. Every restriction should have a corresponding instruction. If you're going to tell a student that may not use AI to write their conclusion, you also have to teach them how to write a conclusion. Otherwise, the policy uh is just a locked door with nothing behind it. Uh so I instruction. Does this policy teach thinking or does it restrict tools? T trainings, the last one. Who trains the adults and when? The question that determines whether everything above is real, not one training, not a side deck in August that nobody remembers by October. Ongoing, protected, funded time for teachers to actually work through what this looks like in their subject with their students. If your school's AI policy does not have a training plan attached and with dates on it, you do not have an AI policy. You have an intention, and intentions don't survive uh contact with a Tuesday in November. Ask your school leader directly when a teacher trained on this, how often and who's paying for the time. If the there's no answer, you found a gap, and it's the one that determines everything else. So, how to use the audit? Run all five, score each one honestly, solid, weak, or missing entirely. Most of the policies I've reviewed, uh, including any early drafts on my own, uh scores well on assumptions and detections and fail badly on usability and training. That's the pattern. Now, now you know where where to look first. So here's what I'm doing and what I got wrong. So, and I want to be honest with my own building, uh, including the parts I didn't get right the first time. Uh, we have an AI policy. Um, we wrote it uh before anyone required us to do it, uh, which puts us ahead of most. But I want to be careful not to sound like I've solved this because I didn't, uh nobody has. Uh, our core line is simple that a teacher can hold. Did AI help you think? Or did AI think for you? Students can use AI to check grammar, to brainstorm, to ask what's weak in an argument. They cannot use it to produce finished work and submit it as their own. And here's what I got wrong initially. My first version was too detailed. It it tried to anticipate every scenario, and the result was a document uh that was technically thorough and practically useless. My teachers couldn't apply it in real time. They failed the U in my uh own framework, and I didn't see it until the teacher told me plainly uh that she stopped consulting it because that wasn't time. Uh, that's how I learned the usability question. She taught, she taught it to me. Uh, so what do we do now? We run regular in-class assessments without devices, um, not as a trap, as information. Uh, when a student sees the gap between their AI-assisted work and their independent work, that's what the most of the teaching moment available. It's honest data and it belongs to them. We train teachers on an ongoing basis, uh, not once, and we treat policy as a living document that we expect to revise because Justin Reek is right, and we're at the 1905 stage of this. Um, anyone who tells you they have a permanent answer is selling you something. And we teach critical thinking as its own discipline. Uh that's the whole premise of uh the thinking to think, this show. If a student can evaluate information, detect manipulation, think through consequences, and slow down before uh before reacting, AI becomes a tool uh they use without those skills. Um it becomes the critical thinking is the foundation. AI is is the building. You cannot build without the foundation. I am created, I I have created, excuse me, um, in an AI tool uh to help and assist teachers. I I started doing consulting on the side for uh implementation such as this. So if you are a school leader and is interested, um, you can contact me. The link is in the description, and uh we can uh set up a an appointment uh to go over your policies, procedures, and see what I can do to assist. I do professional development training for teachers uh outside my own my own school. So you can reach out if you need assistance on this matter. Uh everything will be FERPA protected. For those of you that are not school leaders or in education, it's kind of like HIPAA. Uh and for those outside of the US that are listening, uh, it is uh basically our privacy laws. Um FERPA compliant, um and I can train you on that as well and how to uh navigate that. So, your homework. Depending on who you are, here's your assignment this week. If you're a parent, email your school and ask one question: Does our school have an AI policy? And can I read it? Then run the five audit question against it. If your state is holding hearings and many are right now, uh that's a public meeting you're allowed to attend. If you're a teacher, uh run the usability test in your school's policy. Picture a second period or fifth. Uh, can you apply it in four seconds without stopping instruction? If not, bring that to your administrator's attention. Not as a comp not as a complaint. Please don't do that. You always try to come up with solutions when you're gonna complain. My advice to especially my younger teachers out there, my young educators, don't go to an admin strictly complaining. Go with possible solutions as well. Be part of the problem-solving solution. And so, again, not as a complaint, as a design problem. Um, that's a conversation a good leader wants to have. If you're a school leader, run all five audit questions against your current draft before you adopt it. Be honest about training, especially. Please be honest about training especially. If there are no dates on the calendar, yeah, you have an in an intention, not a plan. And again, you can always reach out to me. I do professional development training on this matter. I have created programs and software. Um, I'm actually piloting uh three uh softwares in the school uh that I will be releasing for uh possible beta testers as soon as I have more data and feedback from teachers as well as students and parents. Um, if you're in Florida, the State Department of Board of Education workshop is always fit. Uh, it concerns a rule that would cover public and charter schools. Know what's being proposed before it's decided. And everyone, ask the question that matters more than any policy. In your own life, in your own work, is that tool helping me think? Or is it thinking for me? Because the answer to that question is the whole thing. And it applies to all of us, not just the students. So uh we covered a lot. Uh, 27 states are actively legislating AI in schools. Um, five are already in law. Ohio is the first hard deadline that just passed. California released a model policy, and Florida State Board uh takes up a rule on August 5th uh that covers charter schools like mine. That's a real movement, and states deserve credit for it. Uh, we also named the four things most of these policies get wrong. They regulate uh tools instead of teaching thinking. They assume detection works when it doesn't. They don't fund the training, and they confuse having a document with having a uh a change. And I gave you the audit framework, assumptions, usability, detection, instruction, and training. Five questions that tell you about in 10 minutes whether a policy will change a classroom or sit in a binder. I'll leave you with this. Policy is necessary, policy is not sufficient. A mandate produces a document, leadership produces change, and the gap between those two is where every kid in every classroom actually lives. Follow the show on Spotify or Apple Podcasts. Um, the one tap is the single most useful thing you can do to keep this conversation going. Um, please comment, uh question. Um, if you don't have any comments or anything, you just want to give a thumbs up saying great or whatever, that's wonderful too, because that actually feeds the algorithm and it does help out the podcast. I will not be doing videos like this often, but in one of my LinkedIn profiles, a commentator made a comment uh believing that my podcast and my my essays are all AI created and something. And I was like, all right, here I am. Um again, the autoframework worksheet, all five questions with scoring, plus a version written specifically uh for parents to bring to a school board meeting is free at uh maponti. Uh my name is literally on the screen, um, but with an additional A. Uh Substack.com. The link is below in the description. If you're a school leader who wants help building or reviewing an AI policy for your organization, consultation information is in the description. Think clearly, lead boldly, stay logical. I'm Mike Aponte, also known as M Aponti, and this is thinking to think. Take care.