By Dr. Janette Camacho | April 22, 2025

In the two years since ChatGPT catalyzed the AI-in-education conversation, the federal government has produced executive orders and white papers but no binding K-12 guidance. That means the regulatory landscape - the rules that determine how 50 million American students interact with AI in school - is being shaped state by state, framework by framework, in a patchwork that ranges from visionary to nonexistent.

As of April 2025, I have reviewed AI education guidance documents from 38 states and the District of Columbia. The variation is staggering. Some states have published comprehensive, actionable frameworks that address pedagogy, ethics, privacy, and professional development. Others have issued one-page memos that amount to "be careful." And several states - including some of the largest in the nation - have published nothing at all.

This article compares the leading frameworks, identifies what separates effective guidance from performative gestures, and argues that every state owes its educators more than silence.

The Leaders

California: Comprehensive but Complex

California's Department of Education released its "AI Guidance for California Schools" in October 2024, followed by a more detailed implementation toolkit in February 2025. At 87 pages, it is the most thorough state document I have reviewed.

Strengths: California's framework explicitly addresses algorithmic bias, requiring districts to evaluate AI tools for disparate impact on historically marginalized student populations. It includes a decision-making matrix for administrators evaluating AI vendors, specific guidance on AI and students with disabilities, and a professional development planning template.

Limitations: The document's length and complexity may limit adoption in smaller districts without dedicated technology leadership. Several administrators I spoke with described it as "useful but overwhelming." California also stops short of mandating AI literacy standards, instead recommending their adoption - a distinction that matters in a state with 1,037 school districts and enormous local autonomy.

North Carolina: Practitioner-Centered

North Carolina's "AI in NC Public Schools" framework, released in stages throughout 2024 and consolidated in early 2025, stands out for its practitioner focus. Rather than leading with policy language, it organizes guidance around teacher use cases: lesson planning, assessment, differentiation, and communication.

Strengths: Each section includes annotated examples of effective and ineffective AI use. The framework explicitly names specific tools (ChatGPT, Gemini, Copilot, Khanmigo) rather than speaking in abstractions, which makes it immediately applicable. North Carolina also established a statewide AI in Education Advisory Council with practicing teachers comprising at least 50% of membership - a structural commitment to practitioner voice that I find exemplary.

Limitations: The privacy guidance is less detailed than California's, and the framework does not address procurement standards for AI tools in the same depth.

Oregon: Equity-First Framing

Oregon's approach, formalized through its 2024 "AI for Oregon Learners" initiative, is distinctive for centering equity throughout the document rather than treating it as a separate section. Every recommendation includes an equity impact analysis asking: Who benefits? Who might be harmed? What data supports this?

Strengths: Oregon's framework includes specific guidance for tribal communities, rural districts, and schools serving high percentages of English learners. It addresses the digital infrastructure requirements for AI adoption, acknowledging that tools requiring high-bandwidth internet access create access barriers in rural communities. This is the only state framework I reviewed that explicitly grapples with the material prerequisites of AI integration.

Limitations: Oregon's guidance is strong on principles but lighter on implementation specifics. Teachers have told me they appreciate the "why" but want more of the "how."

The Middle Tier

A significant number of states - I count roughly 15 to 18 - have published guidance that is well-intentioned but insufficient. These documents typically share several characteristics.

They focus on risk mitigation rather than pedagogical opportunity. The predominant tone is cautionary: do not let students use AI to cheat, protect student data, be aware of bias. These are important messages, but a framework that addresses only risk creates a culture of avoidance rather than a culture of thoughtful integration.

They lack specificity. Statements like "districts should develop AI acceptable use policies" are common. But without model policies, decision frameworks, or implementation examples, this guidance asks districts to solve a problem the state itself has not solved.

They ignore professional development entirely. Of the 38 state documents I reviewed, only 11 include substantive guidance on teacher training. The rest assume that teachers will somehow acquire AI competency through osmosis or personal initiative. This is the most consequential omission. You cannot implement a framework through a workforce that has not been trained to understand it.

They are static. Most state guidance documents were published once and have not been updated. In a field where the technology changes quarterly, a document from early 2024 is already outdated in important ways. Leading states (California, North Carolina) have committed to regular revision cycles. Most have not.

The Laggards

As of this writing, at least seven states have published no AI-specific guidance for K-12 education. I will not name them individually because the list changes as new documents emerge, and I do not want to penalize a state that may be days from publication. But the pattern among lagging states is consistent: they tend to have smaller state education agencies, fewer dedicated technology staff, and less political pressure from organized educator advocacy.

The cost of state-level silence is borne by districts. When a state provides no guidance, each of its school districts must independently research, develop, and implement AI policy. This means that affluent suburban districts with dedicated technology directors and legal counsel produce sophisticated frameworks, while rural and under-resourced districts - the ones that most need support - produce nothing or produce poorly informed policies based on fear.

State-level inaction is, functionally, a decision to let wealth determine AI governance in education. I find that indefensible.

What Every Framework Should Include

Based on my analysis and my experience implementing AI integration across diverse school settings, I believe an effective state AI framework must include seven elements.

1. A clear position on AI's role in education. Is AI a tool to be leveraged, a threat to be managed, or both? States need to articulate a vision, not just a set of restrictions.

2. Grade-banded guidance. AI use by a second grader and AI use by a twelfth grader are fundamentally different. Frameworks must differentiate expectations by developmental level.

3. Specific privacy standards. Which student data may be processed by AI tools? What vendor agreements are required? How should districts evaluate AI tools against FERPA, COPPA, and state privacy laws?

4. Academic integrity guidance. Not blanket prohibitions, but nuanced frameworks - like the CLEAR model I described in my July 2024 article - that distinguish between productive AI use and academic dishonesty based on context and learning objectives.

5. Equity impact requirements. Mandatory consideration of how AI policies affect students with disabilities, English learners, economically disadvantaged students, and students in low-connectivity environments.

6. Professional development mandates. Not suggestions. Mandates. With funding attached. Teachers cannot implement what they do not understand.

7. A revision schedule. Any framework published in 2025 will need updating by 2026. States should commit to annual review cycles at minimum.

What Educators Can Do

If your state's framework is strong, advocate for your district to adopt and implement it fully. If it is weak or nonexistent, you have two immediate options.

First, borrow from the leaders. California's, North Carolina's, and Oregon's frameworks are publicly available. Adapt them to your context. You do not need to start from scratch.

Second, organize. State legislators and state board members respond to constituent pressure. Parent groups, teacher unions, and professional organizations can all advocate for comprehensive AI guidance. The states that have produced strong frameworks did so because organized stakeholders demanded them.

The AI policy landscape in American education is fragmented, uneven, and - in too many places - empty. That is not acceptable. Our students deserve coherent guidance, our teachers deserve professional support, and our communities deserve transparency about how AI is being used in their schools. The states that recognize this are building frameworks worth emulating. The rest need to catch up.

Dr. Janette Camacho is a Google for Education Certified Trainer & Coach, Google Certified Educator Level 1 & 2, Adobe Creative Educator, Apple Teacher, FETC 2024 and 2025 Featured Presenter with 28+ years of K-12 classroom experience. She is the founder of iTeachAI.