Illinois publishes a school AI guide that does not require using it

On July 9, 2026, the Illinois State Board of Education (ISBE) published statewide guidance to help school districts decide how to use — or not use — artificial intelligence in K-12 education. The document responds to Public Act 104-0399 and does not require classrooms to adopt AI tools. Its purpose is to offer a framework for choosing them with judgment, protecting students, and keeping teachers and communities at the center of the decision. The timing is part of the news: the guidance arrives with the 2026-27 school year on the horizon, just as districts decide which tools to buy and which policies to write.

The most striking detail of the story is that the initial drafts received help from AI tools, according to Capitol News Illinois's coverage. But that transparency does not change the substance: the guide does not delegate education policy to a machine. It was developed by ISBE with a panel of specialists and requires each district to justify the purpose of any use and to keep human oversight.

Who wrote it and who signs it

The official release puts names to the advisory panel, and the mix is itself a statement of intent. It includes Julia Wynn, senior director of State Government Affairs at CodeAI, from industry; Rebecca Winthrop, director of the Center for Universal Education at the Brookings Institution, from education research; Tina Halliman, superintendent of the SPEED Special Education Cooperative District 802, from school practice with the most vulnerable students; and James Larimore, co-founder and chairperson of the EdSAFE AI Alliance, the organization promoting safety standards for educational AI. Industry, academia, special education, and standards: four different chairs at the same table.

The institutional position is summed up by State Superintendent Tony Sanders in a sentence that works as the whole document's thesis: "AI is a tool to support teaching and learning, not a replacement for the relationships that ground teaching and learning."

A guide, not an adoption mandate

ISBE explains that the guidance creates no state mandates to use AI. Instead, it gives districts practical resources, model policies, and cases so they can adapt their decisions to local context, capacity, and community priorities. Nor is it a list of recommended products: no brands are blessed or banned, and that absence is deliberate too. Guidance that does not mandate can end up a dead letter or become the de facto standard; training and follow-through make the difference, which is why the 2026-27 resource plan is the part that will decide whether this document changes classrooms or just shelves.

The difference matters. One administration may decide a tool helps prepare materials, support accessibility, or ease an administrative task. Another may conclude it is not ready, for lack of training, privacy agreements, or evaluation mechanisms. A good guide does not remove that decision: it forces it into the open.

The document rests on four ideas. Teaching and learning are grounded in human relationships; schools serve academic, developmental, and civic purposes; AI informs teaching rather than replacing it; and its use demands deliberate purpose and local oversight. It is a sober formulation against two common simplifications: that banning everything solves the problem, or that introducing a tool is in itself innovation. Each of the four also cuts off a different shortcut: the first dismantles the fantasy of the artificial tutor replacing the classroom; the second recalls that school also teaches how to live together and decide — a tool that optimizes only grades ignores two thirds of the mission—; the third denies AI the role of oracle; and the fourth returns the decision to the ground, with deliberate purpose and local oversight instead of a uniform policy dictated from above.

What it asks districts to review before choosing a tool

The guidance covers instructional planning, digital citizenship, privacy, security and legal compliance, along with equity, accessibility, bias mitigation, governance, and procurement. It also recommends that staff, families, and the community agree on why and how AI will be used before selecting products and deciding how implementation will be tracked.

That order matters. A district that buys a platform before knowing which educational need it is trying to solve risks measuring success by usage, not learning. Likewise, a tool that seems useful to one teacher may be inappropriate if it processes student data without sufficient guarantees or leaves out those without equal access. The presence of a special-education superintendent on the panel is not decorative: students with specific needs are at once those who can gain the most from accessibility tools and those most exposed if their sensitive data is mishandled.

From the guide's topics comes, almost by itself, a list of questions any family or teacher can bring to their school's next meeting: what pedagogical need does the tool solve; what student data goes into it and under what privacy agreement; who reviews what it produces and who answers when it errs; what happens to students without the same access at home; and how success will be measured — by learning, not by usage. None requires knowing about neural networks. All require that someone thought before buying.

ISBE insists that AI should support, not substitute, the relationships that sustain learning. That does not mean every decision must be made without technology. It means a tool's output needs professional judgment, context, and a responsible person who can explain why it was used.

The value of declaring how it was drafted

That an AI guide acknowledges AI assistance in its early drafts has a limited but real value: it turns transparency into a practice, not just a recommendation aimed at students. The important question is not whether generated text appeared at some stage of the work, but what human review it received, which sources support the decisions, and who answers for the final document. The distinction the guide models with its own example is the one that matters in any organization: using AI to produce a draft is not the same as using it to make the decision. The first is a work tool one declares; the second, a delegation of responsibility that no transparency fixes.

For schools, the lesson is broader. Policies should not stop at asking "was AI used?". They should clarify what it was used for, what information went into the tool, which part a person reviewed, and how an error gets corrected.

ISBE plans to expand resources, role-specific guidance, and free professional learning during the 2026-27 school year, and maintains a dedicated portal for the guidance with an email address for district questions. The challenge will be getting that help into concrete decisions: a purchase, an activity, an assessment, or a communication with families. The guide does not solve every case, but it proposes a discipline that can last: decide the educational purpose first and the technology second. It is the same discipline that works outside school, and that is the capability a reader can take from Illinois even an ocean away.