When a class reads I denne verden er alt mulig, students have easy access to chatbots that will happily write a summary or an analysis on request. Teachers are right to be concerned, but banning tools outright rarely works and can push use underground. A clear, realistic policy helps students understand what is acceptable and why. It also makes it easier to address problems when they arise.

The first step is to decide what you are protecting. If the goal of the assignment is to develop the student's own interpretation, then outsourcing the thinking defeats the purpose. If the goal is to practice revision, then using a tool to get feedback may be perfectly fine. Writing the purpose of the assignment into the instructions makes the policy feel logical rather than arbitrary.

Be specific about what is allowed. Students are often unsure whether using AI to brainstorm, check grammar, or get feedback counts as cheating. A short list of permitted and prohibited uses removes the guesswork. It also reduces the chance of unintentional violations.

Designing assignments that resist shortcuts

Assignments that depend on class-specific discussion, personal response, or process documentation are harder to outsource. Asking students to reference a particular in-class conversation or to submit a short outline and draft history makes it more difficult for generic text to pass. These features also encourage genuine thinking. They are good pedagogy independent of AI concerns.

• Require a brief planning step or outline submitted before the essay
• Ask for reference to specific in-class discussions or activities
• Include a short oral check-in or reflection on the essay's choices
• Collect drafts at several stages rather than only the final version
• Define clearly which AI uses are permitted and which are not

Responding when you suspect misuse

Detection tools are unreliable, and relying on them alone risks accusing honest students. A better approach is to talk to the student about the essay and ask them to explain their reasoning or revise a section in front of you. Students who wrote the work can usually do this easily. Those who did not often struggle, which gives you more reliable information than a software score.

Approach these conversations with curiosity rather than accusation. Some students use tools in ways they did not realize were prohibited, and others feel pressure they have not expressed. A calm discussion often reveals the underlying issue. It also preserves the relationship, which matters for future learning.

Using AI transparently as a teaching tool

One way to reduce the temptation to hide AI use is to model transparent use yourself. Show students how a tool can provide feedback on a draft, then demonstrate how you decide what to accept and what to ignore. This teaches critical evaluation of AI output. It also reframes the tool as a resource rather than a shortcut.

Revisit the policy as tools change. What is realistic this term may be outdated next year, and students notice when rules do not match reality. Invite student input when updating guidelines. Policies shaped with their participation tend to earn more respect.