Fair Process When You Suspect a Student Misused AI

You are an experienced academic-integrity lead. Help me respond fairly to suspected AI misuse.

The assignment and the AI rules students were given: {{ASSIGNMENT_AND_RULES}}
What makes me suspicious, described factually (for example a style change, content beyond what we taught, invented references): {{CONCERNS}}
Evidence of the student's normal work I can compare against: {{COMPARISON_EVIDENCE}}
School or exam-board misconduct procedure (paste the relevant steps): {{PROCEDURE}}

Please:
1. Sort my concerns into stronger and weaker indicators and explain why. Note innocent explanations, such as help from home, a tutor or genuine improvement.
2. Explain plainly why AI-detector scores are unreliable and shouldn't be the only evidence.
3. Plan a calm, non-accusatory conversation where the student explains their process, choices and sources, and can talk through or redo a section.
4. Set out what to record, who to involve and which steps the pasted procedure requires, especially for exam-board coursework.
5. Suggest proportionate outcomes for different findings, from "no concern" to "follow formal procedure".

Constraints: assume nothing until evidence is gathered. Where the pasted procedure is silent, say "check with your exams officer or leadership".

Output: an evidence table, the conversation guide, then the record-keeping and next steps.

Privacy: I've removed names and identifying details. If anything could still identify a real person, tell me to remove it and don't repeat it.