Student AI Use & Integrity

Share / assign

Why it works

AI detectors are unreliable and produce false positives, sometimes flagging honest work, especially from multilingual students. The more durable response is to define, per task, what AI use is acceptable and what crosses into doing the thinking for the student, teach that explicitly, and design tasks where the thinking is visible.

The research: Research on AI-detector reliability and false positives; guidance on academic integrity.

The run-of-show

Choose your slot. The agenda, timings, and length update to match.

The core activity: set the AI line for one task

Teachers leave having set a clear AI line for one assignment and redesigned it for visible thinking.

Pick an upcoming assignment.
Decide what AI use is acceptable for it and what is not, specifically.
Plan how you will teach and communicate that line to students.
Redesign one element so the student's own thinking is visible: a draft, an in-class step, an explanation.

Teach the line, do not police itClear, task-by-task expectations plus assignments that make thinking visible beat an arms race with unreliable detectors.

Facilitator notes

Set expectations per task, not one blanket rule.
Be explicit and teach the line openly.
Design for visible thinking, and do not rely on detectors.

Do not accuse based on a detectorAI detectors produce false positives and can wrongly flag honest students. Never base a serious consequence on one.

Adapt it

Elementary: focus on honesty norms and simple expectations.
Secondary: task-specific AI policies and visible-thinking design.
All settings: teach the line openly rather than hiding it.

Participant handout

One page for every teacher. Print it, or save it as a PDF.

K12 Academics · Professional Learning

Student AI Use: draw the line, teach it

Clear and task-specific: what AI use helps, what crosses the line, and designing for visible thinking.

The assignment:
What AI use is acceptable and what is not:
How I teach and communicate the line:
How I make thinking visible:
How I avoid relying on detectors:

Make it stick

Line per task, thinking visible:

Set the AI line for each task.
Design assignments so thinking is visible.
Bring a task with a clear AI line to the next PLC.

Common questions

Can't I just use an AI detector?

Detectors are unreliable and flag honest work, including from multilingual students. Do not make high-stakes decisions based on them. Teach the line and design for visible thinking instead.

Where is the line between help and cheating?

It depends on the task and should be stated for each one. Using AI to brainstorm or check may be fine; using it to produce the thinking you are being assessed on is not.

Should I just ban AI?

Blanket bans are hard to enforce and leave students unprepared. Clear, task-specific expectations, taught openly, work better.