California Drew a Legal Line. Your State Probably Didn't.
As of fall 2026, California has enacted the first state-level restriction specifically targeting how AI can be used to deliver K-12 instruction. The law does not ban AI from classrooms. What it restricts is narrower and more important: AI operating as the primary instructor, delivering core instruction to students without a credentialed teacher in the oversight role.
That distinction matters enormously, because almost every controversy about AI in schools collapses the two things together. A teacher using AI to draft a quiz is not the same as a student receiving a lesson from an AI system with no human checking the output or the learning. California's legislature decided those are legally different situations. Most state legislatures have not decided anything at all.
According to Stateline, lawmakers and districts across the country are still scrambling to set up guardrails as AI use in schools accelerates. According to Gallup News, most teachers receive no formal guidance on AI use. The Transparency Coalition, a 501(c)(3) nonprofit, issued a legislative update on August 7, 2026, characterizing students as being "treated as test subjects in a massive, unregulated experiment" with generative AI. That is advocacy language, not a finding, but the underlying observation, no consistent standards, rapid deployment, unclear accountability, holds up across multiple sources.
Before you walk into a back-to-school night or sign an enrollment form, you need a working framework for what AI roles actually exist and what questions to ask.
The Four-Tier Framework: What AI Is Actually Doing
Not all AI use in schools is equivalent. These four tiers have meaningfully different implications for your student's learning and for who is accountable when something goes wrong.
Tier 1: AI as Grading Tool
AI scores multiple-choice assessments, flags potential plagiarism, or surfaces patterns in class-wide performance data for a teacher to review. The teacher makes the final call. This is the least disruptive tier and the most defensible from an oversight standpoint. The cognitive work still happens between student and human instructor.
Tier 2: AI as Tutoring Supplement
A student asks an AI tutor to explain a concept they didn't understand in class, practice math problems, or get feedback on a draft essay. A teacher assigned the work; AI helps the student work through it. This tier is common in online and hybrid programs. The risk here is drift: if the AI tutor becomes the de facto teacher because the human teacher is unavailable or unresponsive, Tier 2 has quietly become Tier 4.
Tier 3: AI as Lesson Planner
Teachers use AI to generate lesson plans, differentiate materials for various reading levels, or build assessments. Students may never interact with the AI directly. The teacher reviews, edits, and delivers the content. This tier primarily affects teacher workload and content quality, not student-AI interaction.
Tier 4: AI as Solo Instructor
An AI system delivers the lesson, responds to student questions, assesses understanding, and moves the student through content, with a human checking in periodically or not at all. This is what California's law targets. The cognitive and oversight differences from Tiers 1 through 3 are significant: no credentialed teacher is reading the room, catching the student who's confused but not asking, or making the judgment call that this student needs a different approach today.
The problem is that districts rarely label which tier they're operating in. A school can describe its program as "AI-enhanced learning" whether it's Tier 1 or Tier 4.
State-by-State Snapshot: August 2026
Regulatory status on AI instruction varies sharply by state. Here's where things stand as of this month, based on available legislative tracking.
- California: Enacted legislation restricting AI as primary instructor in K-12 settings. First state to draw this specific legal line.
- Texas, Florida, New York: No enacted legislation specifically governing AI's instructional role as of August 2026. Some district-level policies exist but are inconsistent.
- Oregon, Illinois, Delaware: Legislative activity tracked by White & Case LLP's AI regulatory monitor, but no enacted K-12 AI instruction standards as of this writing.
- Most other states: Silent. No legislation, limited district policy, and, per Gallup News, most teachers in these states have received no formal guidance on what AI use is or isn't appropriate.
The practical result: if your student attends school in any state other than California, the limits on AI's instructional role depend entirely on your district's internal decisions, and most districts haven't made them explicitly.
This is also the dynamic that makes the choice of school, including online school, more consequential than it might appear. Families exploring options like Concord Online High School or Davis Online High School are, by geography, already operating under California's framework. Families in other states, from League City Online High School in Texas to Lake Oswego in Oregon, are operating under whatever the school decided, which may or may not be documented anywhere.
Decision Tree: What Role Is AI Playing in My Student's School?
START: Does your student interact directly with an AI system during school hours?
│
├── NO → Ask: Does the school use AI to grade work or build lesson plans?
│ ├── YES → Tier 1 or Tier 3. Ask for the teacher review policy.
│ └── NO → AI use is minimal or undisclosed. Ask why they haven't answered.
│
└── YES → Who assigned the AI interaction?
├── A teacher assigned it as supplemental practice → Tier 2
│ Ask: How does the teacher review what the AI told your student?
└── The AI system IS the lesson delivery → Tier 4
Ask: Is a credentialed teacher reviewing progress and available for questions?
├── YES, daily → Document it. Ask how.
└── NO or 'periodically' → This is what California's law restricts.
You are now deciding whether you accept that arrangement.
The Checklist: 8 Questions to Ask Your School Before Fall
Print this. Bring it. Don't accept "we use AI responsibly" as an answer to any of them.
- Which AI systems does your school use, and what specifically do they do? (Grading? Tutoring? Lesson delivery? All three?)
- Does a credentialed teacher review AI-generated content before it reaches students? (Every time, or by exception?)
- When my student interacts with an AI tutor or tool, how does their teacher know what happened? (Is there a log? A summary? Nothing?)
- What is your policy on AI delivering instruction when a teacher is absent or unavailable?
- Has your district adopted a written AI use policy? Can I read it? (If the answer is no, that is the answer.)
- How do you verify that AI-generated instructional content is accurate? (Hallucination is a documented problem. Who catches it?)
- Can my student opt out of AI-delivered instruction and receive instruction from a human teacher instead?
- Does your AI use comply with your state's current guidance? (If they say yes, ask them to name the guidance document.)
We applied a similar checklist approach to Kansas's phone ban policy earlier this year, same principle: ask for the specific policy document, not the talking point.
The Bottom Line Before You Enroll
California's law is the first legal precedent that treats AI instruction differently from AI assistance. That precedent will expand to other states or it won't, but either way it gives parents a concrete standard to reference when they walk into a district office. "Does your AI use meet California's standard on instructor oversight?" is a question any school should be able to answer, regardless of where it's located.
The Transparency Coalition is right that students are currently in an unregulated experiment, in the sense that standards are absent and deployment is fast. That doesn't mean AI in education is automatically harmful. It means the accountability structure that would catch harm when it happens isn't built yet.
You are not going to build that structure. The school is. Your job is to find out whether they have.
Families exploring fully structured online options with documented human teacher involvement can start with Bakersfield Online High School if you're in California, or Kissimmee Online High School if you're in Florida. The checklist above applies regardless of where you land.