Scaffolding AI use across K–12
The judgment students need in high school begins with habits developed years before they use an AI tool.
Schools often ask when students should start using AI. An equally important question is what students need to know and be able to do before they begin.
Asking how we know something, comparing accounts, identifying sources, and recognizing contradictions all contribute to evaluating AI output. Students also need subject knowledge: they cannot reliably assess an explanation they do not understand.
The progression below gives those foundations time to develop. It proposes deliberate stages of introduction, with direct use following classroom evaluation and increasing responsibility.
Kindergarten through third grade: build the foundations
In this approach, kindergarten through third grade involves no student AI use or dedicated AI lessons. The emphasis is on foundational knowledge and habits of inquiry.
Children learn to ask how someone knows something, compare differing accounts, explain where an idea came from, and notice when a statement conflicts with what they have observed. They also practice attempting a problem, explaining their approach, and revising their thinking with support.
These habits belong in ordinary instruction: discussing a picture book, observing a plant, comparing historical accounts, or explaining a mathematical answer. They have value throughout the curriculum and will later help students examine AI-generated material.
Protect time for this work. Students need opportunities to sustain attention, encounter uncertainty, and work through difficulty with appropriate help. Immediate completion should not become the measure of a successful lesson.
Fourth and fifth grade: examine output together
Introduce AI-generated material as something the class can examine, with the teacher operating any approved tool.
The class might pose a question and compare the response with a text, observation, or other source. Students discuss what the answer gets right, what it leaves out, and which claims need checking.
The lesson should not depend on catching the tool in an error. An accurate response also invites useful questions: What supports it? Is the explanation complete? Does it answer the question we asked?
Students' work remains evaluation, comparison, and explanation. They do not need individual accounts to begin learning that fluent language is insufficient evidence of accuracy.
Sixth through eighth grade: introduce supervised use
Where product age requirements and school approval permit, students begin using tools themselves within clearly defined tasks. Teacher supervision includes reviewing their decisions and checking practices, as well as managing access.
Build a short routine into the work:
- Before use: What help do I need, and is it appropriate for this task?
- During use: What should I question, and how can I check it?
- After use: What did I accept, change, or reject? What did I contribute?
Teachers should also check what students learned. A polished answer does not show whether a student can explain the reasoning or apply it independently.
This is an appropriate stage to practice disclosure on routine assignments. Teach students to name the tool, describe the assistance, and explain what they did with the result. Our guide to what students should disclose about AI use provides examples.
Students who are not ready or eligible for direct access can continue evaluating teacher-provided material while developing the same habits.
Ninth through twelfth grade: assess independent judgment
High school students should have opportunities to select among approved tools, decide whether assistance is useful, evaluate responses, and account for their choices.
The central question is whether assistance supports the learning or performs the thinking the student needs to practice. That distinction depends on the assignment. AI-assisted research may be appropriate for one task; another may require students to find and evaluate sources independently.
Ask students to justify their decisions and demonstrate the underlying knowledge. They should be able to explain their argument, support their claims, and identify limitations in the assistance they received.
Independence should increase as students demonstrate readiness. Grade level alone does not establish it. Our guide to AI literacy describes the capabilities this progression is intended to develop.
What carries through
Across the grades, students return to a common set of questions:
- How do we know this?
- Where did the information come from?
- What evidence would change my mind?
- What did I contribute, and what help did I receive?
The material becomes more demanding, and students assume more responsibility for answering. A discussion about a picture book and an evaluation of a generated research summary draw on related habits, though the later task also requires deeper subject knowledge and an understanding of AI's limitations.
Schools should check both as students progress. Early work in inquiry and explanation provides a foundation; later instruction must teach students how to apply it to the tools and tasks they encounter.
Building a K-12 progression?
Our AI literacy framework sets out expectations by strand from kindergarten through graduation, and our training works with teachers on what that looks like in their own grade level.