1. Start with the official scope
Collect the syllabus, learning objectives, assessment format, marking guidance, and instructor announcements. These are the constraints AI cannot safely invent for you.
Use AI to help group the scope into topics and prerequisites, then check the result against the original documents before building a schedule.
2. Use AI to create structure, not certainty
Ask for a breakdown of the material, a comparison table, possible misconceptions, or a sequence of learning tasks. Treat the output as a draft that helps you organize—not as an authoritative version of the course.
- Keep source documents available for verification
- Flag claims, calculations, and quotations that need checking
- Remove generated detail that falls outside the assessment scope
3. Generate practice, then inspect it
AI can quickly create questions at several levels, but quantity is not the same as quality. Check whether each item has a defensible answer, uses the right terminology, and matches the kind of thinking the exam requires.
Whenever possible, include instructor-provided examples and past papers. They offer better evidence of format and difficulty than a generic prompt can.
4. Diagnose gaps instead of collecting content
Answer before asking AI for an explanation. Then compare your reasoning, identify the earliest point where it went wrong, and request a smaller explanation or related practice for that specific gap.
Update the study plan from these results. Topics you can retrieve and apply need less time; repeated errors and missing prerequisites need more.
5. Protect the skills the exam will measure
If the exam requires writing, calculation, analysis, or judgment without assistance, part of your preparation must reproduce that condition. Put the AI away and complete timed, unsupported attempts.
Follow your institution's academic integrity rules. Do not upload restricted material, personal data, or confidential assessments, and do not submit generated work as your own when that is prohibited.
- Use AI after an attempt when possible, not before every attempt
- Verify high-stakes information with trusted course sources
- Practise regularly without AI so the skill remains yours