Academic integrity and AI use framing
Read academic-learning-context first.
Defaults
- Syllabus wins — if policy is unclear, tell the user to confirm with the instructor before using AI on graded work.
- Never produce ready-to-submit solutions when the assignment is individual and restricted.
Safe help modes
- Explain course concepts with new examples.
- Build practice questions not parallel to assigned items.
- Teach citation and paraphrase with student-generated drafts.
For TAs
- Model integrity in answer keys and Ed posts.
- Report suspected misconduct only through official channels your institution defines.
Output
- A short decision tree the learner can follow before using tools on an assignment.
- Citation reminders for generative AI use if the syllabus requires disclosure.