January is often the first realistic point in the year when schools can pause, look up, and decide whether their AI policy still matches practice. By the end of December, many teams have tested new tools, improvised workarounds, and discovered gaps between written rules and daily behaviour. That makes the start of term ideal for a reset. Instead of treating policy as a compliance document to file away, schools can use January INSET to align expectations, clarify red lines, and set a calmer direction for the months ahead. If you want a companion checklist before the session, the annual AI acceptable use policy refresh checklist is a useful starting point.
A January policy review has one major advantage: staff have fresh examples in mind. They can remember where AI genuinely saved time, where it created uncertainty, and where existing guidance was too vague to help. A September launch often feels theoretical; January feels concrete. Teachers can point to report comments, planning prompts, revision materials, parent communications, or safeguarding concerns they actually encountered.
There is also a practical leadership benefit. A short, structured INSET session can prevent a slow build-up of mixed messages across departments. If one team allows AI drafting with disclosure, another bans it entirely, and a third uses tools without agreed approval routes, inconsistency becomes the real risk. A policy sprint gives everyone the same baseline.
If your last policy review happened even six months ago, the landscape has shifted. Governance expectations are clearer, but the tools are also more capable. Schools now need to think beyond simple chatbot use and address longer conversations, memory-like features, multimodal inputs, and agent-like workflows that can complete several linked tasks with limited supervision.
For many leaders, procurement and governance questions have become more pressing because of the phased implementation of the EU AI Act and its knock-on effect on vendors, contracts, and documentation. Even schools outside the EU are feeling this through supplier terms, product design, and assurance expectations. The EU AI Act governance playbook offers a helpful overview of what this means in practice.
At the same time, DfE guidance for 2025 has pushed schools towards a more mature position: neither blanket enthusiasm nor blanket prohibition, but proportionate use with strong human oversight. That means your policy now needs to say more about who approves tools, how outputs are checked, and what evidence is retained when AI contributes to important work.
Model behaviour has changed too. Newer systems can maintain longer context across extended chats, making them more useful for planning and administration, but also more likely to accumulate sensitive detail over time. The briefing on extended workflows and governance is worth reading if your staff increasingly work in long-running conversations rather than one-off prompts.
A January INSET policy sprint does not need to be elaborate. In fact, it works best when it is tightly timed and visibly practical.
Begin with 10 minutes on what has changed since the last review. Keep this focused on three things: regulation, platform capability, and school experience. Staff do not need a lecture on AI history. They need a clear answer to: “Why are we updating this now?”
Use the next 20 minutes for role-based discussion. Ask classroom teachers, middle leaders, safeguarding leads, SEND staff, data protection colleagues, and senior leaders to review the policy from their own perspective. A teacher might ask whether AI can help draft model answers. A safeguarding lead might ask what happens if a pupil discloses harm to a public chatbot. A middle leader might ask how departmental resources created with AI should be labelled and quality-assured.
Then spend 25 minutes on clause review. Put your current policy on screen and work through a short list of priority sections: acceptable uses, prohibited uses, approval routes, data protection, assessment integrity, safeguarding, and accountability. The aim is not to perfect every sentence in the room. It is to agree the direction, identify wording to update, and assign owners.
The next 15 minutes should focus on scenarios. Give staff short cases drawn from reality: a teacher uploading a pupil paragraph for feedback, a head of year using AI to draft a parent email, a pupil using an image generator for homework, or a leader relying on AI summaries from a long meeting transcript. Ask, “Allowed, not allowed, or allowed with conditions?” This is where policy becomes usable.
Use the final 20 minutes to confirm next steps, publication dates, and staff support. If you want a ready-made structure for the wider session design, the INSET day AI workshop guide pairs well with this policy sprint approach.
Every 2026 school AI policy now needs a few core elements. First, it must define approved, restricted, and prohibited uses in plain language. Staff should not have to infer the rules from abstract principles alone. Second, it must state that accountability remains human, even when AI supports drafting, summarising, or planning. Third, it must explain what data can never be entered into unapproved tools.
It should also cover record-keeping. If AI contributes to high-stakes work such as reports, behaviour communications, assessment materials, or safeguarding-related administration, schools need a proportionate audit trail. That does not mean documenting every low-stakes prompt. It does mean being able to show who used which tool, for what purpose, under what approval, and with what review.
The most useful policy clauses are short enough to read and specific enough to apply. Here are examples you can adapt for your own document.
For staff: staff may use approved AI tools to support planning, administrative drafting, and resource generation, provided they review all outputs for accuracy, bias, tone, and suitability before use.
For pupils: pupils may use AI only in ways explicitly permitted by their teacher or school policy, and must not present AI-generated work as wholly their own where disclosure is required.
For leaders: leaders must ensure that any AI tool used at whole-school level has been reviewed for safeguarding, privacy, contractual, and operational risk before deployment.
For everyone: no user may enter confidential personal data, safeguarding records, or assessment-sensitive information into an unapproved system.
Ready to Revolutionise Your Teaching Experience?
Discover the power of Automated Education by joining out community of educators who are reclaiming their time whilst enriching their classrooms. With our intuitive platform, you can automate administrative tasks, personalise student learning, and engage with your class like never before.
Don’t let administrative tasks overshadow your passion for teaching. Sign up today and transform your educational environment with Automated Education.
This is the section many older policies miss. Long-context tools can hold a thread across many interactions, which makes them powerful for curriculum planning, report drafting, and project support. It also means a single chat can gradually collect names, patterns, concerns, and strategic detail that no one intended to store together.
Your policy should therefore state when a new chat must be started, what kinds of information must never be added to an ongoing thread, and whether chat history should be disabled by default where possible. Staff need to understand that risk can build slowly over time, not only through one dramatic mistake. The privacy audit checklist can help teams review retention, export, and deletion settings before problems arise.
Agent-like tools need separate treatment. If a system can search files, draft emails, connect to other apps, or complete multi-step tasks, approval should be stricter. The policy should say which integrations are permitted, who authorises them, and what evidence is captured when such tools are used in operational workflows.
Good policy depends on platform controls. If your school uses tools inside managed ecosystems such as Microsoft or Google, your written rules should match the admin settings actually in place. That includes age restrictions, chat history, file access, sharing permissions, and logging. Otherwise, staff hear one thing in training and experience another on the platform.
Approval routes should also be visible. A simple flow works well: classroom experimentation with approved tools may sit at department level, but any tool handling pupil data, generating reports, or integrating with school systems should require central sign-off. The minimum viable AI toolkit rollout guide is useful here because it connects policy language to operational deployment.
Assessment, safeguarding, and data protection still need the clearest boundaries. In assessment, schools should specify when AI support is prohibited, when it is permitted with disclosure, and how staff should respond if misuse is suspected. In safeguarding, the policy should state that AI tools must not replace established reporting routes, designated safeguarding lead judgement, or direct action where there is a risk of harm.
For data protection, the key principle is simple: if you would hesitate to place the information on a public website, do not place it into an unapproved AI tool. Staff appreciate direct wording more than legal jargon. The challenge is not usually bad intent; it is speed, convenience, and unclear boundaries.
The fastest way to create policy overload is to write a document that sounds comprehensive but feels disconnected from school life. Staff buy in when policy solves real problems. That means showing how the updated guidance protects them as well as the school. A clearer rule on report drafting, for example, removes uncertainty. A defined approval route for new tools prevents staff from carrying the risk alone.
It also helps to acknowledge that AI use exists on a spectrum. Some colleagues need support with first steps. Others need tighter boundaries because they are already experimenting extensively. A strong January INSET session makes room for both groups without making either feel judged.
The month after INSET matters more than the session itself. Publish the revised policy quickly, ideally with a one-page summary for busy staff. Confirm approved tools and switch off anything that no longer meets your standards. Ask line managers to revisit one practical scenario in team meetings so the policy stays alive beyond the launch.
Within 30 days, review whether staff understand the new expectations. Look for repeated questions, inconsistent practice, or hidden friction points. If necessary, issue a short clarification note rather than waiting for the next annual review. Policies improve through use, not just through drafting.
A January AI policy sprint works because it treats governance as a professional habit, not a one-off event. Done well, it gives staff confidence, leaders oversight, and pupils stronger protection as tools continue to evolve. For many schools, that is the most valuable way to begin the year.
Here’s to a clearer, calmer start to your 2026 AI planning. The Automated Education Team
Microsoft Build 2026 brought a fresh wave of Copilot announcements, but …
Summer term is often the last realistic window for school leaders and AI …
Schools do not need a whole-school AI rollout to learn what works. In many …
Once the papers are marked, many departments want feedback that is sharper …
Primary assessment week can feel intense, especially when SATs and spring …
Exam week often creates pressure to do more, faster, with less time to …
Many school AI rules still rely on vague disclosures such as “I only used …
AI voice tools for modern foreign languages have improved sharply by 2026, …
Spring assessment season puts pressure on teachers, pupils and families to …
Jaanuar on praktiline aeg kooli AI vastuvõetava kasutuse poliitika ajakohastamiseks, eriti pärast …
Januari är ett praktiskt tillfälle att uppdatera skolans policy för acceptabel användning av AI, …
Tammikuu on käytännöllinen hetki päivittää koulun AI:n hyväksyttävää käyttöä koskeva käytäntö, …