Why do the old integrity rules feel insufficient?
Traditional rules assumed that work was either the student's own or copied from someone else. AI creates a middle ground: a student may brainstorm with a tool, ask it to check clarity, or have it write whole paragraphs. Whether each of these is acceptable depends on the task, and students cannot guess what you intend. Integrity now starts with saying it out loud.
How do you set clear expectations for each task?
State the rule on the assignment itself. Short, concrete wording works best:
- No AI: "This task must be done without AI tools."
- Limited AI: "You may use AI to brainstorm and check clarity. You may not use it to write sentences you hand in."
- Open AI: "You may use AI throughout, and you must show how."
Rules set task by task are easier to follow than a single blanket rule. For the broader picture, see Can students use AI for schoolwork?
Why does transparency matter?
When students declare how they used AI, teachers can see their thinking and respond to it. Make the declaration safe: an honest declaration of permitted use carries no penalty. If students fear being punished for telling the truth, they will hide their use, and the school loses its best source of information.
Ask for a short statement, such as: "I used AI to ___. I changed ___. I checked these claims in ___." For wording options, see How should students cite or acknowledge AI use?
What counts as process evidence?
Process evidence shows how the work came to be, not only what it says. Useful forms include:
- Prompt logs: the prompts the student used and the responses they received.
- Drafts: early versions that show how the work developed.
- Source checks: a record of the real, non-AI sources used to verify claims.
- Short conversations: a few minutes in which the student explains their choices.
Teachers can then assess the thinking and not only the product. When a written final product is risky, a short conversation or written reflection of equal weight is a sound alternative.
Why is guessing from the final text unreliable?
Student writing varies, and AI writing can look ordinary. AI detection tools can be wrong in both directions, flagging honest work and missing generated work. A hunch or a detector score is not evidence. See Can teachers tell if a student used AI? and Do AI detectors work?
Evidence of process, and a respectful conversation with the student, give a far fairer picture.
What should happen when something goes wrong?
Start from learning. Talk with the student privately, ask them to walk you through how the work was made, and listen. Explain the expectation and why it exists. Then agree on a way to show the learning, such as redoing part of the task with a conversation. Follow your school's policy for any further steps. A full approach is set out in What should a teacher do when a student used AI when they were not allowed to?
Note that unauthorized AI use and plagiarism are related but different. See plagiarism vs. unauthorized AI use.
How does Endorse It connect to integrity?
The fourth pillar of the WISE framework, Endorse It, says that the work AI helps you produce is yours, and you are accountable for it. This is the heart of integrity: whatever tools were used, the student puts their name on the result and can explain it. Process evidence and short conversations are simply ways of making that accountability visible.
Where should you start?
Pick one upcoming assignment. Write the AI rule on it in one sentence, add a short declaration line, and require one piece of process evidence. A free student handout and sample policy are available through the email signup.
Quick answers
Should students be penalized for declaring AI use?
No, not when the use was permitted. An honest declaration of permitted use carries no penalty, because transparency is what makes fair assessment possible.
Can I use an AI detector as proof?
No. Detection tools can be wrong, so a result should never be the only basis for an accusation. Use process evidence and a conversation with the student.