A high school student’s AI permissions belong to the task, not to the grade level. The same ninth grader who pressure-tests an argument with a chatbot on Tuesday sits down Thursday to a ten-minute in-class paragraph with nothing but a pen. That paragraph is the unaided sample everything else she hands in gets read against. Grades 9-12 are where the distinction becomes teachable: older students can carry the whole discipline — recording, verifying, disclosing, defending, transferring — and be assessed on it.

Graduating autonomy task by task

Students at this level should be able to tell six levels apart on sight: AI-free assessment, mechanical assistance, feedback and coaching, research support, AI-integrated disciplinary inquiry, and prohibited substitution. A teacher’s job is to decide which one a given assignment is and to say so in the instructions. Teaching students to make the same sort for themselves is work a blanket rule cannot do.

Blanket rules fail in both directions, and they fail quietly. A department that permits AI across the board gets students using it on the one task that was measuring independent proficiency. Nobody notices until a secure sample comes back well below the take-home work. Ban it outright and the same students use it at home, unsupervised, on the assignments that matter most to them, where nobody at school ever sees it.

Whether a tool fits depends on what the student already knows and can judge. A source-bounded argument coach that sharpens an AP history student’s reasoning may be a poor fit for a fourth grader, who does not yet have the background knowledge to catch it when it is wrong. That is the developmental question inside the Five Tests, and it runs task by task at every level. The elementary and middle school guidance answers it with tighter defaults, recommended by design; no experiment has established an age threshold behind them.

What older students are accountable for

Five obligations make AI-present work assessable at this level. Students maintain process records, verify claims, disclose material assistance, defend their work, and demonstrate independent transfer.

Each one exists because a finished product cannot show who did the thinking. Paper does not settle it either. A blue book gives you a time-bounded independent sample and a reduced ability to consult a chatbot during the period, but a student can prepare material beforehand, memorize it, and reproduce AI-written sentences by hand. A handwritten final answer reveals almost nothing about how sources were chosen or how an argument developed.

The five sort into the same product, process, and performance evidence any AI-present assessment plan rests on: the essay is the product, the annotations and revision memo are the process, the defense and transfer task are the performance.

Ask for documentation only where AI materially affected the intellectual product. Requiring a full transcript for every trivial use can turn a thinking exercise into compliance work. A disclosure worth reading is short: the tool, the purpose, the stage of the work, one material suggestion the student accepted, one they rejected or corrected, and how they checked accuracy. That last line can make the disclosure itself an exercise in judgment.

Setting AI policy one assignment at a time

The six levels a student learns to tell apart are the same distinctions a teacher makes when setting the policy for a particular assignment. The AI and K-12 hub collects the frameworks for making that call. The Four Modes of AI Assignments sets how much AI a given assignment allows, at the assignment and phase level. A schoolwide statement that AI may be used responsibly is too vague to guide behavior. The Five Tests decide whether to authorize the use at all. Attempt, Coach, Verify, Transfer — the sequence for an assignment that opens the tool only after an independent attempt — fixes when in the work it appears. Designing assessment for an AI-present classroom covers the product, process and performance evidence the judgment rests on. The Four Modes, the Five Tests and that sequence are our own synthesis of the research rather than published standards.

Social studies shows the whole shape in one assignment: knowledge and an initial argument built with no AI, a coach bounded to the rubric and source packet that questions but never writes, then a defense and an independent transfer task. The full six-phase sequence, including the Uncommon Schools result behind it, is worked end to end on the social studies page.

Social studies runs the lateral-reading routine NCSS has connected to classroom ChatGPT use — extract a claim, leave the chatbot, verify independently. Embedded teaching like that is usually stronger than a separate lesson about AI. Teaching AI literacy to students covers the student-competency picture.

Naming the level before you set the assignment

The next time an assignment lands where AI could plausibly help, the useful question is which of the six levels that particular task belongs to. If you want more guidance of this kind, the Kindred K-12 newsletter sends practical material for humanities teachers.