An AI policy for schools sets the principles: what artificial intelligence may be used for, what it may not, and what protections apply to students and their data. An AI operational handbook turns those principles into daily practice. It tells a teacher which tools are approved, tells a principal what to do when a student submits AI-written work or an AI-altered image circulates, and tells the technology office how a new tool is reviewed before anyone signs a contract. A district can adopt a sound policy and still leave every one of those questions to individual schools. The handbook is where the district answers them once, for everyone.
The need is plain in the classroom. Teachers use AI to draft lesson materials and differentiate assignments, students use it at home for homework and writing, families ask what is allowed, and AI features now arrive inside platforms districts already license. Yet in a Gallup survey published in May 2026, only 18 percent of teachers said they receive formal guidance on AI from their administrators. Most teachers are making these decisions on their own.
This article sets out what an AI handbook does that a policy statement alone does not, what the evidence says about district readiness, who should write the handbook, and what a first edition should contain. For the wider picture of AI in schools and colleges, see edCircuit’s AI in Education: A Guide for Schools and Colleges.
What Is an AI Policy for Schools, and Where Does a Handbook Fit?
In most districts, policy is adopted by the school board and changes slowly. It states values and limits: student data is protected, academic honesty is expected, staff remain responsible for decisions about students. That stability is useful, but a board policy cannot list which chatbot a seventh-grade class may use this semester or describe the steps a counselor follows when a manipulated image of a student appears online. Those details change too often, and they belong to the people who run schools.
An operational handbook carries that detail. It sits beneath board policy and alongside the acceptable use, student privacy and code of conduct documents the district already has. Districts that have published one show the shape. Houston ISD’s AI Guidebook, developed with AI for Education in 2024 for students, staff and families, sets guiding principles and age-based limits on access, names approved enterprise tools, requires parental consent, and asks users to cite AI-assisted work and verify what a tool produces. In August 2026, Springfield Public Schools in Missouri released its first AI handbook, a 36-page version 1.0 that takes a guided-use approach rather than a ban and that the district says it will revise after hearing from students, staff and families.
The version number is worth noticing. A handbook is expected to change, and a district that labels the first edition as a first edition is telling teachers and families that revision is part of the plan.
Why School Districts Need an AI Handbook Now
State guidance has spread quickly. According to AI for Education’s state tracker, updated in August 2026, 35 states and Puerto Rico have issued official K-12 AI guidance. State documents give districts a frame, but most stop short of operational decisions. When the Illinois State Board of Education released its statewide guidance in July 2026, it offered model policy resources, case studies and practical tools, and it stated that the guidance creates no statewide mandate for classroom AI use. The choices about tools, training and enforcement remain with each district.
District readiness has lagged behind that guidance. In a Digital Promise survey of 31 districts reported by K-12 Dive in August 2024, only 25 percent had set specific AI policies or guidance, though 61 percent said work was underway. The sample is small and the survey is now two years old, so it shows direction rather than a national rate. A larger picture of operations comes from the Consortium for School Networking. Its 2025 survey of 281 members, most of them district technology leaders, found that 76 percent said district leaders were involved in setting direction for AI in operational areas, but many of those districts were still in early discussion. As GovTech reported from the same survey, only 12 percent described clear direction on oversight, 20 percent had conducted legal or risk reviews, and 21 percent had not addressed legal and risk questions at all. The respondents were IT leaders reporting on network and business operations, not classroom practice, but the gap they describe is the one an AI strategy for school districts has to close: interest and activity without settled rules.
Without a handbook, the same pattern tends to appear. Teachers adopt tools on their own, families hear different messages from different schools, and administrators respond to each problem as it arrives, with no shared basis for the decision.
What Teachers Need From the Handbook
Teachers ask practical questions, and a useful handbook answers them directly:
- Which AI tools are approved, and for which grades?
- May a teacher use AI to draft a test, a rubric or feedback on student work?
- How should AI-generated material be checked before it reaches students?
- What should a teacher do when a student appears to have submitted AI-written work?
- What student information may never be entered into an AI tool?
When these answers are missing, teachers tend to take one of two courses. Some avoid AI altogether, including uses that would save them time or help students. Others use it in good faith in ways that later prove to conflict with privacy law or with a colleague’s classroom rules. Clear guidance lets teachers try new practices within agreed limits, and it tells them the district will stand behind a decision made inside those limits.
Consistency Across Schools and Classrooms
Without shared guidance, a student’s experience of AI depends on the school and the teacher. One class may use AI tutoring and feedback tools well; the class next door may avoid them because the teacher is unsure what is permitted. One teacher may allow AI for brainstorming while another treats any use as cheating. Students in the first group gain skills and support that students in the second do not, and students who move between the two learn that the rules depend on who is enforcing them.
A districtwide handbook gives every school the same starting point. Teachers can still set expectations for a particular assignment, but they do so within common definitions of acceptable use, common disclosure expectations and common consequences. Families then receive one message, and administrators can support teachers rather than adjudicate between conflicting classroom rules. edCircuit’s AI Rules for Schools: A Back-to-School Guide covers how schools can communicate those classroom expectations to students and families.
Protecting Students and the District
AI brings real risks into schools: student data entered into tools without proper agreements, inaccurate or biased output, over-reliance on automated scoring or feedback, unvetted tools adopted by individual staff, material that does not match the curriculum, and student misuse that ranges from plagiarism to impersonation and manipulated images. A handbook lets the district decide its response to each of these before an incident rather than during one. At a minimum it should define:
- Data privacy requirements, tied to the Family Educational Rights and Privacy Act (FERPA) and the Children’s Online Privacy Protection Rule (COPPA)
- The approved tool list and the process for adding to it
- Classroom boundaries for staff and for students, by grade band
- Warning signs staff should report
- Consequences for misuse, consistent with the student code of conduct
- Incident response steps, including who is notified and in what order
Two points deserve particular care. First, decisions with serious consequences for a student, such as discipline, placement or special education eligibility, should not rest on AI output without documented human review. Second, the handbook should be honest about the limits of AI detection software, so that an accusation of cheating does not rest on a detector’s score alone. Incident planning for manipulated images and video is covered in edCircuit’s AI Deepfakes: The New School Safety Threat, and the questions districts should put to vendors before buying are set out in AI Procurement for Schools: 10 Questions to Ask.
Training That Matches the Handbook
A handbook issued without training gives teachers rules without the reasons for them. Many districts have begun to train staff: a RAND report published in April 2025 found that about half of districts had trained teachers on generative AI by fall 2024. That still leaves many teachers untrained, and the Gallup finding that fewer than one in five teachers receive formal guidance from administrators suggests that training and written guidance have not always arrived together.
Teachers do not need to become computer scientists to use AI well, but they do need training that is practical, grounded in instruction, consistent with district policy, suited to their grade level and subject, and repeated over the year rather than offered once. Training tied to the handbook should help teachers understand why each guideline exists, what AI tools can and cannot do, when output can be trusted and when it must be checked, and how AI can support differentiation and scaffolding without replacing the student’s own work. Teachers who understand the reasoning are better placed to apply the handbook to situations it did not anticipate.
Who Should Write a School District AI Handbook?
A handbook written by one office tends to reflect that office’s concerns. A stronger first edition comes from a team that includes:
- Chief technology or information officer, responsible for infrastructure, cybersecurity and data governance
- District technology coordinators and IT staff, who understand the tools, their deployment and compliance
- Curriculum and instruction leaders, who keep AI use aligned with academic standards
- Principals and assistant principals, who carry out the handbook in buildings
- Classroom teachers from across grades and subjects, who know how guidance will meet the school day
- Special education leaders, responsible for accessibility, accommodations and legal compliance
- Library and media specialists, who teach information and digital literacy
- Professional development coordinators, who build training to match the handbook
- Communications staff, who prepare messages for families
- Student representatives, where appropriate, who can describe how students actually use AI
Families should also have a way to comment before and after adoption, as Springfield has committed to do. The team should meet at least quarterly so that the document keeps pace with the tools.
What Should an AI Handbook for Schools Include?
A first edition does not need to answer every question. It should be clear about the questions it does answer and say when it will be reviewed. A foundational handbook typically includes:
- A district vision for AI, linked to its instructional goals
- Plain-language definitions for staff, students and families
- Ethical principles, including human responsibility for decisions about students
- Data privacy requirements under FERPA, COPPA and state student privacy law
- Approved and restricted AI tools, with the grades and purposes each is approved for
- Expectations for staff use in instruction and administration
- Age-appropriate rules for student use, and any parental consent required
- Academic integrity guidance, including how students disclose AI assistance
- A process for evaluating and approving new tools
- Cybersecurity protections
- Professional development requirements
- Expectations for communicating with families
- Incident response procedures
- A review and update schedule
Districts do not have to start from a blank page. The TeachAI AI Guidance for Schools Toolkit, developed by Code.org, ETS, ISTE, Khan Academy and the World Economic Forum and updated in 2025, offers principles and sample language for policymakers and school leaders. CoSN’s AI resources for K-12 leaders include a generative AI readiness checklist and a maturity tool that a district team can use to judge where it stands before it writes.
Keeping the Handbook Current
AI tools change faster than policy cycles, so the handbook needs a set schedule for review, with named owners. The technology office, academic services, the curriculum office and building administrators should revisit it several times each year to update the approved tool list, strengthen privacy expectations as vendors change their terms, add instructional uses that teachers have found effective, respond to new state or federal requirements, and address new risks such as AI-edited video and manipulated images. Each revision should be dated and summarized for staff and families so that everyone knows which version is in force.
A district that writes an AI handbook now will not settle every question about AI in education, but it will decide who answers those questions, on what principles, and how the answers reach every classroom. That work reduces risk, supports teachers who want to use AI well, gives families a consistent account of what their children will encounter, and keeps students in different schools on equal footing. It also prepares students for a world in which knowing how to use AI responsibly is part of being literate.
Related reading
Subscribe to edCircuit to stay up to date on all of our shows, podcasts, news, and thought leadership articles.
More on Artificial Intelligence
