Purpose
This template helps schools define AI-era integrity expectations without over-relying on probabilistic tools. Adapt bracketed sections to your governance process. Legal review is recommended before adoption.
Definitions
'Unauthorized assistance' means using generative AI, humans, or paid services to complete work when the instructor has not permitted it. 'Permitted assistance' includes tools explicitly allowed in the syllabus, such as spell-checkers, citation managers, or disclosed brainstorming aids.
'Detector report' means a probabilistic summary of AI-writing signals. It is not standalone evidence of misconduct.
Instructor responsibilities
Instructors shall publish clear AI-use rules per assignment. No finding of violation may rest solely on a detector score. Instructors must offer students an opportunity to respond with drafts, notes, or other process evidence before major sanctions.
Student responsibilities
Students shall follow published course rules, attribute sources accurately, and disclose permitted AI assistance when required. Submitting work that is primarily machine-generated without permission is a violation.
Review and appeals
[Institution name] provides an appeals path within [number] days of notification. Appeals must be in writing and may include draft history, timestamps, and instructor feedback. Review bodies shall consider detector limitations documented in the report.
Data and privacy
Analyses are stored under student control unless exported for formal proceedings. Opt-in anonymized research use requires explicit consent per submission. Pilot programs must disclose if anonymized submissions may inform model improvement.