AI-use policy
Instead of “AI is forbidden,” write down what counts.
The short answer
For each task, a policy answers four questions: may students use AI, what do they hand in besides the final work, how are the parts weighted, and how can a student explain or appeal. Students follow rules they can see.
A guide from Edudojo, an AI that won’t think for students: its coach asks questions instead. The advice works with or without it.
What the policy needs to say
A process-over-product policy says that evidence of learning includes the path to the work. The final work still matters as a synthesis; it is not enough on its own when AI can produce it cheaply. For each task, the policy answers four questions:
- May students use AI, and must they declare it?
- What do they hand in besides the final work?
- How are the parts weighted?
- How can a student explain their work, or appeal a decision?
Start with a short statement your department can approve, then set the AI rule task by task.
Sample statement to adapt
In this course we assess the work you hand in and how you made it. For this task, you may use AI tools. Say which tool you used and what you used it for. With the final work, hand in your drafts (at least one revised) and a short declaration of how you worked (and what help you used). A polished final file without these parts may be treated as incomplete. If you think a decision about your work is wrong, you can explain how you worked and ask for a review.
What students hand in
- Drafts, with at least one revision after feedback.
- A short note on how they worked: which tools or help, and for what.
- The checks they made: sources looked up, AI claims kept or dropped.
- A short reflection: what they worked out themselves, and what a tool supplied.
- Where the stakes are high, an explanation step: oral, a live question-and-answer, or a timed write-up.
Named hand-ins make the policy workable: students know what to produce, and markers know what to read.
How you might weight the parts: an example
| Component | Example range | What it looks at |
|---|---|---|
| Process and inquiry | 25–40% | Drafts, revisions, questions answered |
| Checking and AI literacy | 15–25% | Sources and AI claims checked |
| Final work | 25–35% | Still marked for quality |
| Explanation or reflection | 10–20% | Can the student explain it? |
This table is an example for a teacher to adapt, not a recommendation. The numbers have no study behind them; set the weights for your course and discipline.
Integrity reviews and appeals
Say whether drafts and notes may be read in an integrity review, and that they come before any detector score as evidence. Give students a way to explain when the record and their account differ: saved activity shows what an account did, not what a student thought.
Edudojo keeps the drafts and the coach chat inside the assignment, and the assigned teacher can read them after submission. The weights above are yours to set.
What we do not claim.
- No grades
- Edudojo does not grade. The teacher judges the work.
Try it this week
Write the four answers for one task, and share them with students before the task starts.
Trying this with a college? See the pilot plan