What kinds of policies should we establish to set boundaries and expectations for AI use?

What the Research Says

What does the research say about policies for setting boundaries and expectations for AI use in schools? After a review of our research collection, we found that what is emphasized is that successful and ethical adoption of generative AI in schools hinges on having clear, actionable, and inclusive policies. That means that policies should do more than regulate, they must guide safe, equitable, and purpose-driven use of AI for teaching and learning. Here are a few findings from the research we reviewed.

1. Policies Must Balance Enablement with Guardrails. For example, the University of Michigan’s GAIA Committee (2023) recommends policy frameworks that enable innovation while safeguarding ethical, legal, and secure AI use. Their guidance includes access control policies (who can use which tools), expectations for transparency (AI use in coursework or grading), and academic integrity frameworks that evolve with AI technologies.
2. Equity, Access, and Student Privacy Are Core Policy Pillars. For example, the ISTE Whitepaper (2024) and Center for Democracy & Technology (2024) both emphasize that bans on generative AI are often ineffective and can exacerbate inequity by limiting access for underserved students. Instead, policies should ensure equitable access to vetted AI tools, enforce data minimization and student privacy protections, and require teacher and staff training to identify and mitigate bias.
3. AI Literacy and Professional Development Must Be Embedded. The EDUCAUSE Horizon Report (2023) argues that without embedding AI literacy and digital responsibility into school policies, the long-term benefits of AI are unlikely to be realized. School, district, and university policies should mandate ongoing professional learning for staff, encourage student development of AI fluency and ethical use, and create safe spaces for trial, reflection, and feedback.
4. Policies Should Align with Broader Curriculum and Assessment Strategy. Assessment is deeply affected by GenAI. According to the AI & Assessments Event Summary (2023), educational systems must develop policies that guide appropriate use of AI in formative and summative assessments, reimagine traditional assessment formats to reflect AI-integrated workflows, and define what constitutes original work in the age of AI augmentation.
5. Institutional Frameworks Help Ensure Responsible Innovation. The Sociotechnical Safety Evaluation Framework (DeepMind, 2023) stresses that AI capabilities should be reviewed within context, involving layers of capability, human interaction, and systemic impact. Policy recommendations include ongoing risk assessments, multi-stakeholder input from teachers, students, and communities, and clear procedures for handling misuse or unexpected outcomes.

Policy Category	Key Elements to Include
Acceptable Use Policy	Define approved tools, use cases, and prohibited AI usage (e.g., deception, surveillance)
Academic Integrity	Redefine plagiarism and originality in light of AI-assisted work
Data and Privacy	Require AI tools to comply with FERPA, COPPA, and data minimization principles
Equity and Access	Ensure equitable access to AI tools; provide offline alternatives where needed
Professional Development	Mandatory training for educators on AI literacy, ethics, and bias mitigation
Assessment Policy	Guidelines for AI use in creating, grading, or submitting assignments
Transparency Guidelines	Require disclosure of AI use in student and staff output

At the time of this writing several (but not all states) have developed guidance, and most recently it was stated at UNESCO Digital Learning Week that less than 10% of schools and universities have established frameworks for AI use and adoption. The table below shows summaries and tagged “forward thinking” positions that were noted as outliers in our review of existing state AI and education policies and guidance documents.

An analysis of these forward thinking elements demonstrated consistency across states in the their inclusion of (1) guiding principles focused on ethical, safe, and responsible use of AI, with language about transparency, fairness, privacy, and human oversight; (2) emphasizing the need for professional learning, suggesting PD opportunities, AI literacy training, or embedded supports to help teachers navigate and model AI use; (3) the positioning of AI as a tool to support personalized learning, student agency, and in some cases, inclusion through UDL (Universal Design for Learning); (4) a strong emphasis on academic honesty, plagiarism detection, and defining appropriate vs. inappropriate use, especially in assessment contexts; and the inclusion of templates, frameworks, or suggested steps to empower districts and local education agencies (LEAs) to create their own context-specific AI use policies.

However, some states go beyond these common elements to include more forward thinking elements. Some of which include the following.

• Colorado, Washington, and Oregon emphasize equity, student well-being, and agency as design anchors for AI use.
• West Virginia, Utah, and Kentucky propose phased levels of readiness or AI integration (e.g., “emerging, developing, sustaining”).
• Connecticut, Wisconsin, and California promote cross-sector collaboration, involving families, community members, and students in shaping AI norms.
• Wisconsin and Arizona align AI with broader state digital transformation strategies, embedding AI within IT literacy and library systems.
• Washington, Indiana, and North Dakota include reflective protocols, questions, or self-assessment tools for schools and teachers.

Across Most States there were clear distinctions between generative AI, predictive AI, and acceptable tools. There were are also differentiated responsibilities for administrators, teachers, students, and families. There was also fairly consistent alignment with FERPA, COPPA, and vendor vetting procedures. Lastly, AI by and large is suggested to be integrated in subjects like English, computer science, and digital citizenship.

Only a few states like Wisconsin acknowledge the role of school libraries and non-classroom learning in AI education. While states like Alabama and Georgia provide more templates or portals rather than substantive policy frameworks, leaving more up to districts. Though some states encourage stakeholder input, very few explicitly integrate student perspectives into policy co-design. Although academic integrity is heavily discussed, assessment redesign for an AI-rich world is mostly under-explored (except in places like Connecticut and North Carolina.

What a National AI and Education Policy Review Signals for the United States

There were four key themes that could be perceived as signals for national policy for AI and Education in the United States. Those include:

1. Most documents aim to support innovation while managing risk.
2. A values-driven approach is emerging, where AI is not just a tool, but rather an opportunity for broader educational equity and transformation.
3. There is a strong bottom-up orientation, where states empower local action with adaptable resources.
4. A few states are positioning themselves as national leaders, especially in linking AI to future-ready learning, well-being, and cross-sector collaboration.

Recommendations for National Alignment

In consideration of these findings and in light of more national AI initiatives emerging from the White House and others we provide the follow recommendations for national alignment to ensure a national promise to our students.

1. Establish a National AI Education Framework to support coherence across states.
2. Promote shared repositories of AI-aligned curricular resources, use cases, and PD tools.
3. Encourage multi-stakeholder involvement, including students, in AI policy development.
4. Embed AI literacy into pre-service teacher education and licensure standards.
5. Invest in assessment innovation to reflect AI-era competencies like creativity, synthesis, and problem-solving.

Table of State AI Guidance and Policy Recommendations for AI and Education

State	Link	Summary	Forward-Thinking Elements
Alabama	https://www.alabamaachieves.org/wp-content/uploads/2024/06/AI-Policy-Template-LEAs.pdf	Structured policy template for LEAs. Covers academic integrity, responsible AI use, data privacy, student disclosure, teacher responsibilities, and includes editable sections for local customization. Focuses on practical implementation and classroom alignment.	Supports iterative policy review, emphasizes equity, encourages local adaptation, and recommends AI literacy and professional development.
Arizona	https://www.azed.gov/ai	Web-based hub offering curated resources, webinars, and guidance on ethical AI use, student protections, and district-level case studies. Focuses on enabling informed local decisions.	Emphasizes flexibility and local innovation, promotes co-creation of district policies, and focuses on capacity-building through professional learning.
California	https://www.cde.ca.gov/ci/cr/ai/	Comprehensive guidance hub focused on equity, teaching, assessment, data protection, and professional learning. Provides resources, exemplars, and strategic support for LEAs.	Anchors policy in equity and UDL; promotes co-design with students and communities, cross-sector collaboration, and reflective piloting rather than rigid mandates.
Colorado	https://www.coloradoedinitiative.org/wp-content/uploads/2024/05/CO-AI-in-Education-Roadmap.pdf	Strategic roadmap focused on values-based implementation. Covers learner agency, equity, human-AI collaboration, and includes action areas like professional learning, infrastructure, and partnerships.	Centers AI use on learner agency and well-being, promotes pilot-and-learn cycles, and supports cross-sector co-design with educators and communities.
Connecticut	https://portal.ct.gov/-/media/SDE/Digital-Learning/AI-Guidance-CT.pdf	Comprehensive guidance focused on responsible AI use, roles and responsibilities, equity, data privacy, and instructional integration. Includes readiness checklists and example scenarios.	Promotes co-creation with communities, integrates PD alignment and AI literacy, offers adaptive use cases, and embeds risk-benefit analysis into planning.
Delaware	https://www.doe.k12.de.us/cms/lib/DE01922744/Centricity/Domain/4/Generative-AI-Guidance-DE-2024.pdf	Framework organized into five domains: governance, teaching, professional learning, IT, and equity. Focuses on safe, responsible, and effective AI use. Offers LEA guidance on AUPs, literacy, and cross-role coordination.	Promotes adaptive governance, aligns IT and curriculum leadership, encourages personalization and inclusive access, and supports systemic PD development.
Georgia	https://www.gadoe.org/Technology-Services/Instructional-Technology/Pages/AI.aspx	Curated resource page offering AI templates, national frameworks, webinars, and planning questions. Supports LEA-level decision-making with no state-level mandates.	Promotes district autonomy, links to national best practices, provides role-specific PD resources, and supports innovation in diverse educational contexts.
Hawaii	https://www.hawaiipublicschools.org/DOE%20Forms/AI%20Guidance/HIDOE-Guidance-AI-Use-Students.pdf	Student-focused guidance on acceptable and unacceptable AI use. Covers academic integrity, disclosure, and educator oversight. Includes example language for school policies.	Promotes student AI literacy within digital citizenship; allows for evolving updates; supports local school-level adaptation and ethical use practices.
Indiana	https://drive.google.com/file/d/1i-155XYqyFR_YnHE6veanG7ZpM1NBPB2N44/view	Instructional guidance focused on responsible AI use, ethical awareness, and practical classroom strategies. Covers AI foundations, sample use cases, and educator responsibilities.	Promotes AI for creativity and critical thinking, integrates ethics and bias education, supports transdisciplinary use, and emphasizes teacher modeling and professional judgment.
Kentucky	https://www.education.ky.gov/districts/tech/Documents/AI%20Guidance%20Brief.pdf	Concise brief encouraging responsible, ethical use of AI with local flexibility. Emphasizes alignment with legal protections and supports AI for content creation, differentiation, and assessment support.	Prioritizes educator judgment and discretion, promotes AI for scaffolding and personalization, and allows room for local adaptation and future revision.
Louisiana	https://doe.louisiana.gov/about/newsroom/news-releases/release/2024/08/28/louisiana-releases-guidance-for-responsible-use-of-artificial-intelligence-in-k-12-classrooms	Promotes ethical, student-centered AI use focused on integrity, personalization, data privacy, and educator discretion. Frames AI as a supportive tool rather than replacement.	Encourages proactive instructional planning, promotes human-in-the-loop oversight, supports equitable access, and aligns AI with broader digital strategy and PD.
Minnesota	https://education.mn.gov/MDE/dse/tech/AI/AIEd/	Educator-focused guidance supporting ethical AI integration into instruction, promoting student agency, equity, and responsible classroom use. Includes practical tools and examples for curriculum design and accessibility.	Centers equity and bias awareness, promotes shared leadership in decision-making, embeds UDL principles, and develops long-term AI fluency through educator support.
Mississippi	https://www.mdek12.org/sites/default/files/Offices/MDE/OTSS/DL/ai_guidance_final.pdf	Policy-oriented guidance focusing on privacy, responsible use, academic integrity, and transparency. Provides a decision tree for local districts to guide adoption and oversight.	Incorporates district-level decision tools, prioritizes ethical use, and supports responsible innovation anchored in student protection and clarity.
New Jersey	https://www.nj.gov/education/innovation/ai/	Comprehensive AI vision promoting human-centered learning, educator development, and safe, ethical deployment. Focuses on stakeholder engagement, equity, and transparent AI use.	Aligns AI use with future-ready skills, prioritizes transparency and community collaboration, and situates AI within broader digital innovation ecosystems.
North Carolina	https://www.dpi.nc.gov/districts-schools/districts-schools-support/digital-teaching-and-learning/ai-resources	Resource-based guidance focused on digital teaching, ethical AI use, and professional learning. Offers toolkits, external resources, and integration strategies for educators.	Promotes sustainable capacity building, aligns AI with digital citizenship goals, and supports adaptive ecosystems through open-access tools and collaborative networks.
North Dakota	https://www.nd.gov/dpi/policyguidelines/north-dakota-k-12-ai-guidance-framework	Comprehensive framework built on six strategic pillars including governance, curriculum, professional development, and ethical use. Emphasizes local control and provides practical tools for implementation.	Promotes systems-thinking, supports varied readiness levels, integrates AI literacy and ethics, and prioritizes cross-sector collaboration and iterative feedback.
Ohio	https://innovateohio.gov/aistrategy	Statewide AI strategy integrates education with broader innovation goals. Emphasizes computer science alignment, career exploration, and equitable access to AI tools in K-12 education.	Links AI education to economic and workforce planning, supports cross-agency collaboration, and prioritizes digital equity and infrastructure for long-term transformation.
Oklahoma	https://oklahoma.gov/education/services/standards-learning/artificial-intelligence–ai–and-digital-learning1.html	Blends AI guidance with broader digital learning strategy. Offers definitions, ethical use guidance, curricular connections, and a ‘Digital Learning Compass’ for classroom application.	Encourages exploration over restriction, supports educator autonomy, aligns AI with curriculum and fluency, and uses metaphorical tools to foster professional judgment and student-centered innovation.
Oregon	https://www.oregon.gov/ode/educator-resources/teachingcontent/Documents/ODE_Generative_Artificial_Intelligence_(AI)_in_K-12_Classrooms_2023.pdf	Provides foundational guidance on generative AI for educators. Emphasizes local decision-making, equity, academic integrity, and teacher autonomy without being prescriptive.	Encourages inclusive policy development, early-stage professional learning, and equitable, student-centered approaches to academic integrity and responsible AI use.
Utah	https://www.utah.gov/pmn/files/1116147.pdf	Defines AI types and introduces a tiered risk classification system. Emphasizes review protocols, clear roles, FERPA compliance, and transparency in AI adoption in schools.	Applies a structured risk framework, establishes accountability metrics, supports iterative review of tools, and promotes coherent, systemic integration aligned with teaching goals.
Virginia	https://www.education.virginia.gov/media/governorvirginiagov/secretary-of-education/pdf/AI-Education-Guidelines.pdf	Framework organized around ethics, equity, privacy, and academic integrity. Emphasizes district autonomy, stakeholder roles, curricular integration, and educator support.	Introduces AI literacy framework, universal design for learning, and interdisciplinary collaboration. Supports continuous policy updates and digital equity efforts.
Washington	https://ospi.k12.wa.us/sites/default/files/2024-07/ai-guidance_foundations.pdf	Grounded in human-centered values and equity, the guidance supports local policy development, emphasizes transparency, data protection, and professional learning.	Promotes human-centered AI, use of reflective tools, UDL integration, equity-driven decision-making, and personalized learning support.
West Virginia	https://wvde.us/sites/default/files/2024/03/30438-WVDE-AI-Guidance-v1.pdf	Outlines AI definitions, use cases, and guiding principles with an emphasis on responsible use, educator readiness, student support, and data integrity.	Provides a phased implementation framework, promotes ethical AI integration in curriculum, supports professional reflection, and encourages local innovation within a state-supported model.
Wisconsin	https://dpi.wi.gov/imt/empowering-lifelong-learning-ai-guidance-enhancing-k12-and-library-education	Centers on lifelong learning, digital fluency, and responsible AI use in K-12 and library settings, with alignment to information and technology literacy standards.	Integrates library systems and information literacy, supports community engagement, promotes equity and student agency, and anchors AI in state-level digital learning transformation.