Who this is for: anyone preparing for exams or class tests, or seeking to solidify their understanding of course topics, who wants to use GenAI tools as effective study aids.
Important: your module guidance / VLE remains the authoritative source, particularly regarding regulations for online or take-home exams. The primary goal of studying is to build your own knowledge and recall; use AI to support your learning process, not to replace it.
This is the first article in a series of blogs entitled ‘Smart revision with GenAI’. In this blog post, we’ll show you how to use a GenAI tool to support your revision in a way that’s helpful, ethical, and aligned with what UK universities generally expect (make sure you always double-check local guidelines, though!). Many students are already using GenAI; we want you to use it effectively. This first post sets out some ground rules; the following posts in the series will demonstrate practical uses of GenAI to help with your revision.
GenAI can be a brilliant study partner. It can tidy your lecture notes, quiz you, create practice questions, and help you catch up if you’ve missed a week. But it’s also fallible, sometimes over-confident, and it doesn’t know what your module actually covered. So you still need to stay in charge. We call this keeping the human in the loop.
Revision is more than re-reading.
You need to retrieve information, not just stare at it.
You need to revisit topics over time.
You sometimes need things explained differently.
You often need practice questions.
A GenAI tool can help with all of those.
Your exam will be based on your module guide, lectures, practical’s and seminars. GenAI can easily wander off.
Say:
“I’m revising for an undergraduate module on [topic]. Please stay at that level and don’t add extra topics.”
If it includes something you’ve never seen:
“Remove [content] this isn’t taught in this module.”
GenAI is not your brain. It’s a helper.
A good workflow is:
Ask the tool for a question on your topic.
Answer it yourself first.
Ask the tool to mark it or improve it.
Compare with your lecture notes.
That way, you’re practising recall, not just reading AI answers.
Many students paste slides or handouts straight into a GenAI tool. That’s common, but not always what your university wants.
Safe practice:
If you’re not sure whether you can upload lecture materials to an external GenAI tool, work from a University that provides access to CoPilot or Gemini (often called an enterprise model), your own notes or describe the topic instead.
You can still:
Paste your own notes,
Describe the session (“we covered enzyme kinetics and Vmax”),
Or ask for questions at your level.
GenAI can “hallucinate,” i.e., make things up or bring in content from elsewhere.
To reduce that:
“Only use the text I’ve given you.”
“If you guess, tell me.”
“Stick to undergraduate level in the UK.”
Then check against your notes or textbook. If it doesn’t match, then don’t revise it.
Worst use: “Explain photosynthesis”, you read the output, and nothing sticks.
Better use:
“Ask me 10 questions on photosynthesis, one at a time. If I get one wrong, explain it simply, and then re-test me.”
That turns GenAI into a revision partner.
Quick example prompts
“Rewrite these notes clearly for an undergraduate student. Don’t add new content.”
“Ask me one question at a time on membrane transport. Wait for my answer.”
“If I am wrong, explain the correct answer and give me a memory tip.”
In the next post, we’ll stop talking about rules and start talking about actual revision moves — things you can get a GenAI tool to do today: generate quizzes, make flashcards, create exam-style questions, or help you catch up.
Next: Part 2: GenAI revision tools you can use right now
In this second post in the blog series entitled ‘Smart revision with GenAI’, we’re going to show you practical, copy-and-paste ways to use a GenAI tool to revise. If you haven’t read Part 1of the series yet, that’s where we set the ground rules. You can go back to Part 1: How to use GenAI for revision (without getting into trouble) and then return here.
This post is the “do this” one.
We’ll cover:
Turning messy notes into study materials
Getting the AI to test you (retrieval practice)
Fixing half-understood topics
Practising exam-style questions (including closed-book)
Data/methods practice (great for biosciences)
Catching up if you missed a session
Revising in groups with AI
Making it accessible (different needs, different formats)
What to do if your tool is limited
Use this when: you left a lecture with chaotic notes.
Prompt:
“These are my notes from an undergraduate biosciences lecture on [topic].
Rewrite them clearly in bullet points.
Highlight key terms.
Make 5 flashcards to test me.
Don’t add new content.”
You can also say:
“Keep the language simple, English is not my first language.”
Use this when: you want to make sure you can actually recall things.
Prompt:
“You are helping me revise. Ask me one question at a time on [topic]. Wait for my answer. Do not show the answer until I’ve tried. If I am wrong or partly right, tell me the correct answer, give me a 1–2 sentence memory tip, and then give me the next question.”
(That “don’t show the answer yet” bit stops AI from over-helping.)
Use this when: you “sort of” get a topic but can’t explain it smoothly.
Prompt:
“This is how I currently explain [concept]. Tell me what’s correct, what’s missing and what’s inaccurate. Then rewrite it for an undergraduate student at my level. Then give me 3 questions to check I now understand.”
Use this when: your university uses short-answer or data-handling exams.
Prompt:
“Write 3 short-answer exam-style questions on [topic] suitable for a UK undergraduate [type of exam e.g. closed/open book] exam. Each should be worth 5 marks. After I answer, mark my response, tell me what was missing to get full marks, and show me a model answer.”
If you have a past-paper question, paste it and say:
“Generate 3 more questions in exactly this style.”
Use this when: your exam asks you to interpret tables, figures or lab results.
Prompt:
“Invent a simple experiment comparing enzyme activity at three temperatures. Present the results in a small table. Ask me 4 questions: (1) describe the trend, (2) explain the science, (3) identify a limitation, (4) suggest an improvement. Then mark my answers as a lab demonstrator would.”
If the data is too perfect:
“Make the data slightly messy and realistic.”
Use this when: you missed a week or a session.
Prompt:
“I missed a lecture on [topic] in an undergraduate biosciences module. Give me: (1) the likely learning outcomes, (2) essential terms to know, (3) 5 self-test questions, (4) suggestions for where to read next.”
Then:
“Now quiz me on those five questions, one at a time.”
Use this when: you’re revising with friends and don’t want it to drift.
Prompt:
“We are 3 undergraduate students revising [topic]. Create 10 questions we can take turns answering. Vary the difficulty. After each answer, tell us the correct answer and a short explanation.”
Or:
“Create a short case study on [topic] for a group discussion. Include 5 follow-up questions.”
This way, AI keeps the group on-task.
Not everyone revises best from dense text.
Prompts:
“Rewrite this explanation in simpler English, then quiz me.”
“Turn this summary into a short script I can record and listen to.”
“Break this process into numbered steps.”
“Give me the same content in bullet points, then in question-and-answer format.”
Great for EAL students, neurodivergent students, or just for days when your brain is tired.
Some free tools won’t let you paste a long PDF or images.
Workarounds:
“I will paste my notes in 3 parts. Treat them as one document.” (then paste part 1/3, 2/3, 3/3)
“I can’t upload the figure. I will describe it — then create questions from my description.”
“Summarise this text to under 500 words so I can paste it again for questions.”
If you’re not in biosciences:
Swap “dataset” → “source extract” (humanities)
Swap “lab method” → “research method” (social sciences)
Swap “pathway” → “theory or model” (education, psychology)
The pattern is the same: make AI ask → you answer → AI improve → you check with notes.
In the third post, we’re going to show you how to make AI revise with you in ways that match how memory actually works - spacing, retrieval, mixing topics, and elaboration. That’s how you make revision stick.
Next: Part 3: Make GenAI work with learning science.
In this third post in the series, ‘Smart revision with GenAI’, we’ll show you how to get a GenAI tool to support the kinds of revision that actually make knowledge stick: retrieval practice, spaced practice, interleaving, elaboration, and metacognitive checking.
If you missed Part 2: GenAI revision tools you can use right now, go there first for the “what can I get AI to do?” list.
This post explains the why.
Because reading feels easy, but remembering feels hard. The research on learning says you should:
Test yourself,
Come back to topics after a gap,
Mix related topics,
Explain ideas in depth,
Check what you don’t know.
GenAI can do all of that with simple prompts.
GenAI prompt:
“Ask me one question at a time on [topic]. Wait for my answer. Do not show the answer until I’ve replied. If I am wrong, give me the correct answer in 3–4 sentences, and then ask a similar question to check I’ve learned it.”
That last bit (“ask a similar question”) is what strengthens memory.
GenAI prompt:
“I have an exam in 14 days. I have these topics: [list]. I am weakest on [x and y]. Make me a 14-day revision plan that: (1) revisits the weak topics at least 3 times, (2) mixes topics each day (interleaving), (3) keeps each day under 45–60 minutes.”
Daily:
“Generate today’s tasks from that plan, including five retrieval questions.”
GenAI prompt:
“Create 10 mixed revision questions for an undergraduate student interleaving: cell structure, membranes, enzymes, metabolism. Vary the question types. Ask them one at a time.”
This stops you from cramming one topic to death.
GenAI prompt:
“Ask me 3 ‘why’ and ‘what if’ questions about [topic] to deepen my understanding. After I answer, tell me if I have linked ideas well.”
Bioscience twist:
“Ask me what would happen if this enzyme failed / this gene was mutated / this pathway was blocked.”
GenAI prompt:
“I’m going to tell you what I think I know about [topic] at undergraduate level. After I’ve written it, tell me: (1) what I’ve explained well, (2) what’s missing for exam-level detail, (3) likely misconceptions, (4) 3 questions to fix the gaps.”
That makes AI a mirror, not just a generator.
We can make AI test you only on your materials.
GenAI prompt:
“Here are my notes. Create 10 questions using only these notes. After I answer each question, tell me which part of my notes the correct answer came from.”
That keeps you aligned with your actual module.
5 mins - “Give me 5 retrieval questions on yesterday’s topic.”
10 mins - “Explain the one I got most wrong at my level; now re-test me.”
5 mins - “Ask me 3 ‘why/what if’ questions.”
5 mins - “Give me 3 mixed questions from across the module.”
2 mins - “Based on today, what should I revise tomorrow?”
In the final post in this series, we’ll show you how to spot when GenAI is going off-track, when it invents things, teaches the wrong version, or gives a different answer to your lecturer, and how to bring it back.
Next: Part 4: Spotting GenAI’s mistakes and staying in control.
In this final post of the series, ‘Smart revision with GenAI’, we’ll show you how to be the human in the loop. You now know how to make GenAI help you revise (Part 2) and how to make it match how memory works (Part 3). However, GenAI is still just a tool; it can go off-topic, sound extremely confident, even when it’s wrong, and sometimes it won’t match what your lecturer said!
This post is about control and confidence.
This happens a lot, especially in biosciences, where there are multiple valid ways to teach a pathway.
What to do:
“Stay at UK undergraduate level.”
“Use these terms: [paste the terms from your notes].”
“Remove anything that might be beyond this module.”
If it’s still different, note it and learn the version your module is using. After all, that’s the one you’re examined on.
Fix:
“Only use the material I’ve given you.”
“If you don’t know, say you don’t know.”
“Label any part of your answer that was a guess.”
Then check your notes. If it’s not there, don’t waste revision time on it.
Fix:
“Explain your reasoning step by step.”
“Show the cause–effect chain.”
“Show the structure–function link.”
This is key for the later stages of your degree.
Fix:
“Here is an example exam question from my module. Create 3 more in this style.”
“Use a 5-mark short-answer format.”
“Mark me as if you were the examiner.”
That way AI is copying your format.
Sometimes AI just gives you the answer straight away.
Fix:
“Do not reveal the answer until I have responded.”
“If I leave it blank, then show me the answer and explain.”
This keeps your revision active.
You can still revise with AI without uploading:
Describe the topic (“we covered osmosis and water potential”)
Paste your notes (not the slides)
Ask it to stick to what you describe
Prompt:
“I can’t share the original slides. I will describe the lecture. Ask me questions based on my description only.”
Clean up notes
“Rewrite these notes clearly for an undergraduate student. Don’t add new content.”
“Make 8 flashcards from these notes.”
Quiz me
“Ask me one question at a time on [topic]. Wait for my answer. Don’t show the answer yet.”
“If I’m wrong, explain and re-test.”
Exam style
“Write 3 closed-book style questions at UK undergraduate level on [topic].”
“Mark my answer and tell me what’s missing.”
Concept repair
“Here is how I explain [concept]. Tell me what’s right, wrong or missing. Then re-explain it at my level.”
Data/practical
“Invent a small dataset from a lab practical and ask me to interpret it.”
Planning
“I have 7 days. Here are my topics. I am weakest on [x]. Make a 7-day, 45-min-per-day revision plan, with retrieval every day.”
AI literacy
“Label any parts of your answer that I should double-check in my lecture notes.”
“Tell me which bits might not be on my module.”
This was the final part of our Smart Revision with GenAI series.
Part 1 showed you how to stay safe and make AI work with you.
Part 2 gave you concrete revision moves.
Part 3 showed you how to make AI support retrieval, spacing and interleaving.
Part 4 (this post) helped you spot mistakes and stay in control.
You don’t have to use every prompt; try starting with just one, usually:
“Ask me one question at a time on [topic]. Wait for my answer.”
…then build out from there. Good luck!