Survey after survey now lands in the same place: the overwhelming majority of high school students — around 85% in recent polling — admit to using AI on coursework. The number itself is no longer news. What matters is the institutional response, and so far it has clustered into three camps.
Three responses, two dead ends
Camp one: detection. An arms race schools cannot win. Detection tools produce false positives that damage trust, and false negatives that reward the savviest cheaters.
Camp two: surrender. "They'll use it at work anyway" concedes the point that school exists to build capability. A student who lets AI do the reasoning graduates without it.
Camp three: redesign. Change what is assigned. If the task can be completed by pasting the prompt into a chatbot, the task — not the student — is the problem.
Detection polices the old assignment. Redesign makes a better one.
What redesign looks like
- Anchor work in local, first-hand evidence AI has never seen: your classroom's data, your community's archive, this week's seminar.
- Make thinking public — oral defense, exhibition, seminar — where understanding is visible in real time.
- Assess the process: annotations, drafts, revisions, reflections.
- Introduce new evidence mid-unit and require positions to change.
None of this is exotic. It is what strong inquiry teaching has always been — the difference is that AI just made it non-optional.
