As artificial intelligence (AI) becomes more and more accessible and part of everyday life, schools have been working to keep up with student use, as well as faculty and administration demands. The conversation in schools, from kindergarten through higher education, is quickly moving beyond “How can schools use AI?” toward “How do we build the data, governance, and infrastructure necessary to use AI securely and ethically?”
A 2024 report from Cambium Learning Group found that 61% of administrators were using AI to interpret or analyze student data, and 56% were using it to manage student records. With this connection to personal data, much of it concerning individuals under the age of 18, a focus on data governance is critical. Successful AI use is not just about choosing a technology; it is about creating a structure for that technology to securely work with an organization’s data. In response, institutions are focused on data modernization and developing formal AI governance structures that address data privacy, academic integrity, research, and administrative uses.
Best practices are arising as institutions across the educational spectrum thoughtfully address AI implementation and use. Below, we’ve highlighted some of the key issues to focus on.
Establish Clear Rules for AI Use
Schools should develop AI policies that apply to students, teachers, administrators, and staff. Policies should explain which uses are permitted, what information cannot be entered into AI systems, when human review is required, and how concerns should be reported. These policies should also be living documents that can evolve as technology changes.
New York City Public Schools’ (NYCPS) guidance is organized with red, yellow, and green scenarios. A “Red Light” prohibits the use of AI in making automated or high-stakes choices, such as those that involve student grading, discipline, academic placement, graduation, promotion, counseling, and building Individualized Education Programs (IEPs) or 504 plans. “Yellow Light” scenarios include using AI to provide preliminary translations of bilingual instructional material (which must then be reviewed by qualified staff), and permitting students to use AI for brainstorming, but then ensuring their final work products are created without AI. “Green Light” options include using vetted tools for administrative tasks such as drafting emails, organizing information, translating non-critical text, and brainstorming lesson structures.
Treat Student Data as a Protected Asset
Schools need to ask about and understand how each AI tool they purchase uses the data that is put into it. Schools should understand what information a vendor collects, where it is stored, who can access it and whether it can be used to train AI models. These questions should be part of technology procurement and vendor reviews—not addressed after a tool has already been deployed.
NYCPS guidance states that before an AI tool can be used with student data, it must pass its Enterprise Request Management Application, a process that evaluates compliance with federal, state, and district privacy requirements. NYCPS requires vendors to disclose their AI capabilities and prohibits the use of student data to train AI models.
Create an AI Approval and Vetting Process
The growing number of publicly available AI tools makes “shadow AI”—the use of unapproved AI tools by employees or students—a challenge for schools. A formal vetting process should evaluate privacy, cybersecurity, data retention, accessibility, and other requirements before a tool is approved, but cannot be so onerous that people find problematic workarounds.
School districts and universities are making these processes collaborative, providing users with defined questions and frameworks to evaluate if an AI tool will meet organizational requirements. For example, the University of Texas at Austin has an Approved AI Tools Decision Matrix that employees are instructed to consult before using an AI tool with university data. If the tool will access sensitive information, it must meet a specific set of criteria, including being managed by the university and having contractual protections governing data sharing.
Train People, Not Just Technology
Policies are only effective if people understand them. Students, teachers, and administrators need practical AI literacy covering privacy, appropriate data handling, misinformation, bias, academic integrity, and responsible use. A recent study of higher education students found that more than half of respondents were uncertain whether their AI use complied with institutional rules, highlighting the gap between having policies and ensuring users understand them.
AI guidance for Meridian Public Schools in Michigan was developed following staff training and discussions, rather than simply being handed down as a set of rules. The district is pairing its approved tools with AI literacy and responsible-use expectations.
Keep Humans in the Loop.
Schools should establish clear boundaries around automated decision-making. AI can identify patterns, summarize information, and make recommendations, but high-impact decisions affecting students or employees should receive meaningful human oversight.
Virginia Tech is taking a “human+machine” approach to admissions. The university has experimented with AI-assisted essay review while retaining human reviewers in the process. The model is designed to use AI to improve consistency and efficiency without removing human judgment from an important admissions decision.
To learn more about how education organizations are modernizing data management for better AI use, check out these resources from GovWhitePapers and GovEvents:
- OECD Digital Education Outlook 2026: Exploring Effective Uses of Generative AI in Education (white paper) – This report examines emerging uses of AI-powered tools such as intelligent tutoring systems, collaborative learning platforms, and AI-supported instructional design. It highlights how generative AI can enhance personalized learning, improve teacher productivity, and support more adaptive education systems when implemented responsibly. At the same time, the report warns that excessive reliance on AI-generated answers may reduce student engagement and critical thinking.
- The Convergence of AI-Cybersecurity in Education, Workforce Development, and Campus Infrastructure (white paper) – AI and cybersecurity are now converging with significant national economic and security implications, driving the need to secure AI systems, rethink incident response, and redesign education and workforce preparation through new Cyber-AI roles and essential skills for all learners and workers.
- A Framework for Powerful Learning with Emerging Technology (white paper) – Grounded in three principles—learner-centered, evidence-based, and skill-building—this framework outlines actionable strategies to promote agency, metacognition, accessibility, critical thinking, creativity, and collaboration.
- EDUCAUSE Annual Conference 2026 (Sept. 29-Oct. 2, 2026; Denver, CO) – Professionals and technology providers from around the world gather to network, share ideas, grow professionally, and discover solutions to today’s higher education challenges.
- EdTech Symposium (Nov. 4, 2026; Columbia, SC) – Hosted by the University of South Carolina, this event brings together higher education professionals from South Carolina, North Carolina, and Georgia, along with industry partners and guest speakers, to explore innovative technologies in AI, data, infrastructure, teaching and learning, security, and institutional IT services.
Search GovWhitePapers and GovEvents to find even more information on AI use in education.
