Skip to main content
← All resources

Guide · Teachers

AI Academic Integrity Incident Response Guide

What to do when you suspect AI cheating — a step-by-step teacher's guide covering investigation, conversation scripts, documentation, and when to escalate.

Updated · Natalie Gibson

Treat detector flags as signals not proof, review the assignment's AI policy clarity, talk with the student before deciding, match consequences to intent and clarity, document everything, and escalate patterns or disputes to an administrator.

On this page
  1. Before You Begin: The Two Things to Know
  2. Step 1: Identify What You Actually Have
  3. Step 2: Review the Assignment Context
  4. Step 3: Have a Conversation — Before Making a Determination
  5. Step 4: Make a Determination
  6. No violation / clear innocent explanation
  7. Policy was unclear — student acted in good faith
  8. Violation — first occurrence, student takes responsibility
  9. Violation — student denies despite strong evidence, or pattern of violations
  10. Step 5: Parent Communication
  11. Step 6: Document and Update
  12. Conversation Scripts for Common Scenarios
  13. When to Escalate Immediately
  14. Key takeaways
  15. FAQs
  16. Sources

For teachers and administrators. A step-by-step process for handling suspected AI misuse — calmly, fairly, and in a way that holds up.

Before You Begin: The Two Things to Know

1. Detection tools are not evidence. An AI detector flag is a signal worth investigating — not a finding. Students have faced wrongful academic dishonesty charges because teachers treated a detector score as proof. Use detection output as one data point, never as the conclusion.

2. Most AI misuse isn't malicious. The majority of students who violate an AI policy do so because the policy was unclear, because they didn't fully understand the consequences, or because they were overwhelmed and made a bad call. The response should match the reality of the situation.

Step 1: Identify What You Actually Have

Before taking any action, get clear on the evidence you're working from.

Ask yourself:

  • What specifically raised the concern? (Detection tool flag, writing inconsistency, student behavior, all of the above?)
  • Is this inconsistent with prior student work — or with other signals (verbal participation, process documentation)?
  • Could there be an innocent explanation? (Student used AI for brainstorming and the influence leaked in; student is a strong writer and was flagged by a tool that over-flags; non-native English speaker flagged incorrectly)

At this stage: do not contact the student or family yet. Document your observations and move to Step 2.

Step 2: Review the Assignment Context

Check your own records before the conversation.

  • What AI level did you assign for this assignment? (None / Brainstorm / Draft+Revise / Disclosed)
  • Was the policy communicated clearly in writing, or only verbally?
  • Did any student ask for clarification on AI use for this assignment?
  • Is this the student's first concern, or has there been a prior pattern?

If the policy wasn't clearly communicated in writing: Consider whether this is a teachable moment rather than a disciplinary one. Ambiguous policy is an assignment design problem, not purely a student problem.

Step 3: Have a Conversation — Before Making a Determination

A direct, non-accusatory conversation with the student is your most reliable investigative tool and your strongest due process protection.

Set the scene:

  • Private, low-pressure — not in front of peers
  • Framed as understanding, not interrogating: "I want to talk with you about this assignment"

Questions that reveal genuine engagement:

  • "Walk me through how you approached this assignment."
  • "Where did you start? What did you do first?"
  • "Can you explain what you meant by [specific sentence or claim]?"
  • "This word / phrase / argument stood out to me — tell me more about it."
  • "Did you use any AI tools for this assignment? If so, how?"

What to listen for:

  • A student who genuinely wrote the work can almost always explain it — even imperfectly, even nervously
  • Inability to explain specific claims or phrasing that appear verbatim in the submission is a strong signal
  • Unprompted disclosure during the conversation should be treated as a positive, even if the underlying use was a violation

After the conversation, document what was said. Notes taken immediately afterward are far more useful than memory three weeks later.

Step 4: Make a Determination

Based on your evidence, policy clarity, and the student conversation, you're now making a judgment call — not running a detection algorithm.

Likely outcomes:

No violation / clear innocent explanation

  • Document and close
  • If the policy was ambiguous, clarify it in writing for the whole class

Policy was unclear — student acted in good faith

  • Use as a teaching moment, not a punitive one
  • Clarify the policy in writing
  • Consider allowing a redo
  • Document, but no academic integrity record

Violation — first occurrence, student takes responsibility

  • Follow your school's first-offense policy
  • Allow disclosure to reduce consequences if your policy supports it
  • Grade consequence per policy (often a zero or required redo)
  • Parent notification per school protocol
  • Document in student record per school policy

Violation — student denies despite strong evidence, or pattern of violations

  • Escalate to department head or administrator
  • Do not resolve this unilaterally if there's a dispute
  • Follow formal academic integrity process per school handbook

Step 5: Parent Communication

When parent contact is required:

What to say:

I want to let you know about a concern that came up with [student]'s recent [assignment]. Based on our conversation, [brief, factual description of what happened]. Here's what happens next: [consequence per policy]. I want to make sure we're aligned on expectations going forward.

What not to say:

  • "The AI detector caught your child" — it didn't catch anything, it flagged something
  • "I know they used AI" before you're certain — frame as "I have a concern" until you have a full picture
  • Anything that makes this feel like a prosecution rather than a process

Give the parent a chance to share additional context before the conversation ends.

Step 6: Document and Update

After every incident, regardless of outcome:

  • Written record of the concern, evidence, student conversation, and determination
  • Copy of relevant assignment instructions on file
  • Parent communication documented
  • If policy was unclear: update the assignment for next time

One incident often reveals a policy gap. Updating the assignment design or policy language is a legitimate and useful outcome of any AI integrity investigation.

Conversation Scripts for Common Scenarios

Opening the conversation:

I'd like to talk about your recent [assignment]. I'm not here to accuse you of anything — I just want to understand how you approached it.

If the student becomes defensive:

I hear you. I'm not making any decisions right now — I just want to understand your process. Can you walk me through it?

If the student discloses mid-conversation:

I appreciate you telling me that. Let's talk about what happened and figure out the right next step together.

If the student denies and you have strong evidence:

I understand you're saying that. I have some specific things I'd like to walk through with you. [Specific question about a specific sentence or claim.] What can you tell me about this part?

When to Escalate Immediately

  • Student becomes distressed beyond what the situation warrants (consider whether something else is going on)
  • Parent becomes hostile or threatening
  • There are allegations of cheating at the exam or test level (not just assignment)
  • This is a pattern, not a single incident

In any of these cases, loop in your department head or administrator before continuing.

Part of HeyOtto's Teacher's Guide to AI in the Classroom resource series. See also: Why Policy Alone Won't Stop AI Cheating and The Rules of Responsible AI Use.

FAQs

Are AI detection tools enough proof of cheating?
No. A detector flag is a signal worth investigating, not a finding. Use it as one data point alongside process evidence and a student conversation.
Should teachers contact parents before talking to the student?
No. Identify what you have, review assignment context, then have a private student conversation before contacting family or making a determination.
What if the AI policy wasn't clear in writing?
Treat it as a teachable moment: clarify the policy for the class, consider a redo, document the event, and avoid a punitive academic integrity record for good-faith confusion.

Sources

Written for K-12 teachers and administrators handling suspected AI misuse with due process, clarity, and proportional response.

Next step

Want the full product walkthrough?

Request a demo to see how these practices show up in the teacher dashboard and classroom controls.