Guides: AI in schools, activity by activity
Plain answers to the questions teachers, school leaders and parents ask about AI in education. Each one is specific, and each one says where AI should not be used as well as where it should.
- AI in education statistics 2026: what US state AI guidance for schools actually says
The 2026 Compound Intellect State AI Guidance Review: 25 state K-12 AI guidance documents coded against seven criteria, with a per-state table, quotes, method and downloadable data.
- How to turn a school AI policy into classroom practice
A school AI policy sets rules for the whole building; classroom practice happens one activity at a time. Here is a five-step method for turning one into the other, with a worked Biology example.
- Where should AI not be used in a course?
AI should be kept out wherever the student's own work is the evidence: individually assessed work the policy covers, the practical or observation a unit is built on, and closed-book tests. How to decide, and how to record who decided.
- AI lesson planning that follows your school's AI policy
Start from the scheme of work you already wrote, not a blank prompt. In about forty-five minutes a teacher gets a term plan with AI placed activity by activity under the school's policy, plus lessons and editable resources.
- Rolling out AI across a school: a guide for school and district leaders
Set the AI policy once, govern what teachers may propose, start with one department, and judge the rollout by what actually gets taught. Plus the data questions to ask any AI vendor before a pilot.
- The four ways to use AI in a classroom
Every classroom activity falls into one of four AI decisions: Teach with AI, Teach about AI, Teacher-use only, or AI-free. What each means, when to choose it, and an example of each from a real Biology course.
- A free AI policy for schools: the Compound Intellect Standard
The full text of the Compound Intellect Standard, the default AI policy that ships with Compound Intellect. Six permitted uses, two fixed prohibitions, three integrity rules and a one-line student disclosure. Free to read and adapt.
- School AI policy checklist: what a policy needs before a classroom can use it
Eleven things a school AI policy must state before a teacher can plan from it: tools, age, consent, devices, permitted and prohibited uses, integrity, disclosure, assessment windows, data, and who decides.
- Student AI disclosure statement: one line, with examples
A one-line AI disclosure statement students add to their work: “I used [tool] to [what]. I [what I did with it].” Why one line works better than a form, with examples for different subjects.
- How to protect summative assessment from AI without banning it everywhere
Plan the boundary before the work is set: name the assessed activities, mark them AI-free with the authority that requires it, and use AI deliberately in the practice that leads up to them. Detection after the fact is not a plan.
- Where AI should and should not be used in a science course
In science, keep AI out of the practical observation and the assessed write-up, and use it to critique designs, surface misconceptions and audit explanations. A unit-by-unit example from a real Biology course.
- 12 questions to ask an AI education vendor before a pilot
Twelve questions for school and district leaders to ask any AI vendor before a pilot: data and training, student information, export and deletion, subprocessors, teacher monitoring, and where the product keeps AI out.
- How Compound Intellect compares with other AI teacher tools
Most AI teacher tools generate resources one at a time. Compound Intellect plans where AI belongs across a whole term, under the school's own policy, including where it does not. A fair comparison with MagicSchool, Brisk, Khanmigo and the platforms.
- The AI Assessment Scale and the four classroom decisions: how they fit together
The AI Assessment Scale sets how much AI an assessment allows, from No AI to AI Exploration. Compound Intellect's four decisions plan every activity in a term, including teacher-only use and AI as the object of study. They work well together.
- How to teach AI literacy inside the subjects you already teach
AI literacy is best taught inside existing subjects, by making AI output the thing students examine: checking it against evidence, finding its errors and bias, and attributing what it produced. No separate course needed.
- Where AI should and should not be used in a maths course
In maths, keep AI out of fluency practice students must do themselves, timed and closed-book tests, and assessed problem solving. Use it to generate varied practice, explain a worked solution after an attempt, and as a flawed solution for students to correct.
- Where AI should and should not be used in an English course
In English, keep AI out of assessed essays, timed writing and the first draft where students find their own argument. Use it to argue against a thesis the student already holds, give feedback on a draft they wrote, and as a text whose style and claims students critique.
- Where AI should and should not be used in a history course
In history, keep AI out of source analysis students are assessed on, essays and the judgment of evidence. Use it as a flawed summary of a source for students to check, to surface perspectives for discussion, and as an example of how invented citations happen.
- AI with students under 13: how primary and elementary schools can use it safely
With children under 13, the teacher operates the AI tool and students work from its output. Many AI assistants set 13 as a minimum age in their terms, and US law (COPPA) protects children under 13 online.
- Student data privacy and AI tools: what schools need to know
The safest AI tools for schools need no student data at all. Keep names, named work, assessment and safeguarding data out of AI tools, check vendor terms on model training, and know when FERPA and COPPA apply.
- Using AI to adapt resources for learners with additional needs, without naming anyone
AI can adapt one resource for different learning needs (processing time, reading load, working memory, written output, sensory load, getting started) without ever holding a student's name, diagnosis or score.
- AI feedback on student writing: how to use it without replacing the writing
AI feedback helps only after the student has written a draft themselves. Set the order (draft, then feedback, then revision), keep assessed writing AI-free where policy says so, and ask students to disclose what they changed.
- AI at school: a plain guide for parents
What parents should expect from a school's use of AI: a written policy, clear AI-free assessment, no personal data in AI tools, a one-line disclosure when students use AI, and teachers who decide.
- AI in education glossary
Plain definitions of the AI-in-education terms schools meet: AI-free, AI literacy, the AI Assessment Scale, disclosure statements, hallucination, FERPA, COPPA, the four classroom decisions and more.
The pilot is one department, one term, free and with no card. You keep every plan, package and resource whatever you decide afterwards.
Talk about a pilotbasharat@compound-intellect.com