IMPORTANT NOTICE - READ THIS FIRSTIn this section you will find a number of articles on generative artificial intelligence (GAI).
We hope you find the guidance in this section helpful but before following any of the advice on this page or using GAI in any way for your University work, you must read the University of Dundee’s guidance on GAI use for students to ensure you understand the acceptable and unacceptable uses of GAI in assignments.
And remember, even if you ‘only’ use GAI in the brainstorming stage, you need to acknowledge you've used it (see the University’s guidance on acknowledging AI in assessments
You should also make sure there are no School, subject, or assignment-specific guidelines that you need to be aware of.
This page was written by Clara Seyfried (PhD Candidate in Psychology, and ASC Tutor)
Have you ever had a conversation with ChatGPT or a similar large language model (LLM)? Have you ever consulted it because you needed help with your studying or coursework? Then there’s a pretty high chance you’ve already had to deal with a form of prompt engineering, which is a fancy term for the things you have to input in order to get GAI to do exactly what you want. So long as you adhere to the university’s guidelines [INSERT LINK}, you might use AI for any individual aspect of your studying from organising your notes to proofreading essays. However, the output you receive when interacting with an LLM might not always be the most useful. This is where prompt engineering comes in.
In general, the rule of thumb for conversing with a LLM is “the better the prompt, the better the output” – the more you have thought about exactly what you need, the easier it is for the GAI to provide it.
Compare the following interactions with ChatGPT:
a)
When you ask GAI a brief, open-ended question like “how can I improve this”, the most straightforward response to your query might be simply to correct what you have written. But this can be a trap! Rewriting or fixing mistakes generally obscures the exact changes the LLM has made, which means you might lose important information. And, of course, you should typically not copy text (re)written by GAI directly into an assignment for university (and especially not without acknowledging it), as this could raise questions about plagiarism.
Instead, you will get more out of your interaction with GAI if you ask it to give you pointers about what you could improve on, rather than prompting it to fix things itself.
b)
The second prompt is a lot more specific, providing some context about the question, as well as specific requests for the GAI’s output. In response to this query, ChatGPT lists specific feedback about content, style, and language, and ends with the above three suggestions summarising what you might improve about the essay. Note that here ChatGPT picks up on a lot of detail, including the factual inaccuracy of some of the statements and potential style issues, as opposed to just making the language sounds more flowery like in the first example. Though we have shortened it here, the original output original was also much longer and more detailed!
Various frameworks exist that summarise what generally works well for getting the best output from LLMs. One of our favourites, and an excellent one to start with if you’re new to prompt engineering, is the CLEAR Framework.
C – Concise: Be clear about what you want the GAI to do and exclude irrelevant information and superfluous language (e.g., no need to be very polite, just say what you want)
L – Logical: Pay attention to the logical flow of information in your prompt. For example, outline an order of things you want the GAI to cover (e.g., from first to last, by temporal order etc).
E – Explicit: Clearly signal which format, content, and style you want to be included in the output. You can be as detailed as you want (e.g., you could specify exactly how many ideas you want it to generate, or experiment with different output formats).
A – Adaptive: Remember that you can always change your prompt if you’re unhappy with the output (e.g., you could specify your prompt if it fails to answer your question or make it broader).
R - Reflective: Be critical about the output GAI generates and evaluate which elements of it might be useful for you. And of course, you can always go back and prompt the GAI with different requests that might help you more (e.g., ask it to add references to its claims and check if they align with the claims).
This framework already tells you most what you need to know to get high quality output. And often thinking about how to request the information that you need can already help you make a significant step towards finding the right answer! If you are wanting to learn more about how to do effective prompt engineering, check out this article [INSERT LINK TO OTHER ARTICLE].
While GAI can save a lot of time on some tasks, it can be very useful to bear in mind the strengths and weaknesses of how AI functions. As described in the CLEAR framework, you can always adapt your input if you are not happy with the output GAI provides you with. Unlike with a human helper, who might get exhausted or annoyed at you, GAI will never complain - you can always ask for more output and make good critical decisions about which bits of information you find most useful!
On the other hand, one of the downsides of using AI is that it can “hallucinate” information, which is GAI talk for “making things up”. It can also be very biased towards certain perspectives and worldviews. There are ways you can mitigate this (such as prompting it to refer to specific resources or databases) but generally you should not take anything AI tells you for granted, and should always verify information using reliable sources!
Take some time thinking about your instructions – the better the input, the better the output!
The CLEAR framework suggests making prompts concise, logical, explicit, adaptable, and reflective
Beware of AI making things up. That being said, feel free to exploit the wealth of different outputs it can generate!
This page was written by Clara Seyfried (PhD Candidate in Psychology, and ASC Tutor)
Since the term was first coined, the idea of prompt engineering, has garnered much interest on the internet. You can easily find plenty of resources, including prompt libraries (i.e., collections of useful prompts, such as https://library.maastrichtuniversity.nl/apps-tools/ai-prompt-library/ ) that can help you systematically adapt your inputs into generative AI (GAI). But if you are looking to improve the way in which you use GAI, it can also be helpful to think about the different ways in which you can approach your interactions with these large language models (LLMs).
You might already be familiar with frameworks or advice specifying what types of input results in the best GAI generated output [e.g., the CLEAR framework which we looked at in Prompt Engineering – Getting GAI to do what you want: INSERT LINK].
It is important to understand that LLMs function like “black boxes” – we cannot definitively say which prompts result in better outcomes, and determining factors might change drastically over time. As a result, prompts that prevent LLMs from defaulting to their typical behaviours can yield the most interesting results. But whatever you are using GAI for, in many cases it can help you to think about your intentions behind using it, and exactly what you want to get from the interaction, before you start prompting.
Typically, a useful guiding principle for prompting is to think about the role you want the GAI to play in the interaction, the context it needs to understand the task, and the output you are expecting:
Role: This can specify the position of the GAI, including what kind of knowledge it might have and how it should interact with you, your own role in the interaction, as well as the expected nature of the interaction. There are no limits to your imagination! For example, you could ask the GAI to explain a concept to you while pretending to be one of your favourite characters from pop culture or debate a topic with an expert in a very specific field!
Context: The amount of context you provide can have a huge impact on whether the GAI’s output aligns with your expectations. As long as you don’t explicitly state it, the AI might for instance not know that you are a university student in a given field. If a module allows you to use AI, it can be useful to refer to learning outcomes to increase the likelihood that you are engaging with materials you are expected to study.
Output: Depending on what you are using GAI for, outputs might be written text, images, graphs, or files of any given format. However, also think exactly about the style that you want your output to be in. Should it be structured in a specific way? Are there things that should definitely be included?
Many people approach conversations with generative AI like they would human conversations. But just because this is the format in which LLMs are usually presented doesn’t mean that you have to interact with them like you would with a human. Since these models do not necessarily process language in the way humans do, this could even distract you from the best possible ways to prompt them.
Beyond simple input-output prompting (e.g. when you provide instructions or questions to directly receive a response), there are various other ways in which you can prompt GAI models, including:
Chain-of-Thought Prompting: Since GAIs predict linguistic input sequentially, this prompting technique aims to prevent AI from jumping to nonsensical conclusions. GAIs are asked to not only display a specific output but rather to “show their work” by describing each reasoning step before generating a solution.
Self-Consistency Prompting: Since there is never a guarantee that the output generated by GAI is reliable, it can be useful to make GAI check its own work. For example, you might ask it to generate a bunch of outputs and then ask it to rank its answers, providing explanations.
Tree-of-Thought-Prompting: This requires GAI to explore different lines of reasoning at the same time, comparing different viewpoints before arriving at a conclusion. An example for this would be asking GAI to let three experts discuss a topic from different viewpoints.
These are only some techniques are drawn from an overview of prompts that might be useful in higher education
(https://educationaltechnologyjournal.springeropen.com/articles/10.1186/s41239-025-00503-7), so this list is not necessarily exhaustive. Whichever prompting technique works best for you depends very much on the task you are trying to achieve. Experimenting with different techniques can help you find out which ones work best for you!
Considering that the number of ways in which you can phrase your prompts is virtually infinite, it can be good to be creative about ways in which you can incorporate the use of generative AI into your studies. Generally, GAIs are very good at drawing on vast amounts of information and providing information based on specific prompts. This means they are particularly well-suited for scenario-based tasks such as role-playing, as well as providing specific scenarios in which you can test your knowledge (e.g., fictional case studies if you are a medical student). Having ingested most text on the internet also generally makes GAIs a useful tool for certain writing advice, although the usefulness of different prompts always depends on the quality of the prompts [see INSERT LINK TO OTHER ARTICLE].
Depending on the task you use, it can be highly beneficial to provide specific examples of output you want to see. Prompting without examples is sometimes referred to as “zero shot prompting”, but “few-shot prompting” (e.g., introducing a few examples such as relevant input-output pairs) can improve performance markedly. If you are unsure about which prompt to use, remember you can also ask generative AI for advice on which prompt is the most likely to result in output resembling specific examples!
Finally, consider potential ethical risks about working with GAI, especially if you are planning to share personal, confidential, or copyrighted material data with the GAI. And always ensure your use of GAI complies with the university’s guidance. INSERT LINK
Always think carefully about what you want to get out of interactions with generative AI. Specifying roles, context, and output can lead to better prompts
Consider alternative types of prompting that suit your specific needs, including chain-of-thought, self-consistency, and tree-of-thought-prompting
Think creatively about which scenarios generative AI might be most useful (but be careful about which information you share)!
This page was written by Dr Conner McAleese (ASC Tutor and PhD English)
Although GAI tools are not a replacement for academic writing, many students feel using GAI as a digital assistant can enhance the writing process (Kim et al, 2024) in different ways:
GAI may be a productive contributor to planning and structuring academic writing
GAI may help to improve a student’s confidence when learning
This blog post will focus on how you can use GAI to support academic writing. Specifically, this post shall offer you some tips on how GAI can contribute to your ability to form an argument and structure it coherently. Do remember to always check the confines of your assessment and School’s policy on GAI usage before proceeding.
There are many creative opportunities where GAI can be used to support learning and thus improve confidence in a topic. Mike Sharples (in UNESCO, 2023) suggests a range of activities which utilise GAI, one of which is simple question generation using GAI as a ‘Socratic Opponent’ to help make better sense while you learn. Question generation can help you to affirm what you already do know and open up avenues for further investigation for those things you don’t know, yet. Using GAI as a Socratic opponent, will help you to achieve two things pretty quickly. Firstly, you will quickly have an idea whether your question was effective by how well you understand the answer it generates. Using questions effectively is a skill to help build an academic argument in your writing.
Let’s take an example to illustrate how this might work in practice. If I were to ask GAI “What is a mobile phone?” – it’s answer will be pretty generic. Whereas if I were to ask GAI “What would a Roman citizen do with a phone?”, its response will be considerably more imaginative. While these examples may seem silly, they serve to show how important it is to think creatively about how you approach responding to an essay question.
Secondly, you should question the GAI’s response. If we take my example above “What would a Roman citizen do with a phone?”, here is one GAI response:
Clearly, the GAI has attempted humour here, recognising the flippancy of the question. However, from this I have four new areas for exploration in my own speculative essay. For starters, I can see that Baths are important to the Romans. Secondly, Chariot Races appear to be important culturally. Thirdly, I have space to look up and define what a “Taberna” is and the role it played in the day to day lives of Roman Citizens. And, finally, I can assume that a “Gladiator Fight” was also important, culturally to Romans. From a simple – and silly question – I now have a few avenues down which I can proceed to build my argument.
And this where the Socratic Opponent may become crucial to both your own understanding of a topic and structuring an argument around it. GAI works best by generating dialogic conversation. By following the structure of the initial response and continuing to use simple, straightforward, and open-ended questions, GAI can help you to plan an argument around any subject. Follow up questions can probe for more detail, challenge responses, take sides, or simply seek clarification. Through debating and conversing with GAI, there is an opportunity to develop your ideas and add depth and breadth to your academic writing. Always remember to fact check any GAI responses before using them more fully.
Playing with GAI in this way can help build your confidence in a subject by exploring a topic from a range of angles. This in turn will help you to consider how you want to present an argument more coherently.
Some next steps to consider are:
Prior to supplying GAI with your essay question, write down your gut instinct when responding to the question.
Then, have some fun with your gut’s responses and apply creative questioning to GAI.
Once you have some information, ask GAI to offer structure, which you can use as a starting point for your essay.
Kim, J; Yu, S; Detrick, R: and Li, N. (2024) ‘Exploring Students’ Perspectives on Generative AI – assisted academic writing. Available at: https://link.springer.com/article/10.1007/s10639-024-12878-7 Accessed: 28/05/25
UNESCO. (2023) ‘ChatGPT and artificial intelligence in higher education: quick start guide.’ Available from: https://unesdoc.unesco.org/ark:/48223/pf0000385146 Accessed: 02/06/25.
Originally published as a blog post 4th December 2023
This guest post was written by Emma Duke-Williams (CTIL)
Revising, as all students know, can be challenging. It’s difficult to re-read notes, while making sure you understand what they mean, as staff rarely ask you to just regurgitate facts, they want you to show that you understand the content.
So, could Artificial Intelligence tools help you?
Hopefully you’ve seen the Use of Generative AI for Students published in September. This guidance summarises many points about Generative AI (GAI), including a list of (some) potential uses and some points to be aware of.
I’ll start with a few things you should know:
If you’re doing coursework or any graded assessments the University’s position is that you must NOT use GenAI unless its use has been specifically authorised by your lecturer in the assessment brief. This includes remote online exams (i.e. when you’re not invigilated on campus).
However, today we’re looking at your revision. Alongside your highlighter pen and flip cards, could GAI be another tool in your revision kit?
Let’s take some of the ideas for potential use of GAI and think about them in context of your revision.
If you’d rather not share your phone number with external bodies, then we suggest going to Bing.com, signing into it with your University Account and using Microsoft CoPilot. You may, of course, already have an account on ChatGPT, Bard, Claude, etc., in which case you can use that instead.
Once you’ve logged in to your GAI of choice, here are a few potential ways it can help you with your revision.
GAI can summarise longer texts and documents to help you check your own understanding of the key messages and concepts presented.
There are a number of ways to do this, but I’d recommend following Nathan Beel’s advice. He reminds you to check what is generated, and make sure not to use it for your coursework. He also talks about using GAI to generate some multiple choice questions; perhaps you and a friend could both generate sets of questions from two different papers you should have read, and then swap.
There are other tools that will summarise pdfs, such as ResearchRabbit. You’ll have to create an account, and the free accounts are limited to 3 uploads a day.
GAI can act as a conversational or debating partner to develop your ideas and thinking. GAI is very good at answering questions you pose it. If you’re not sure about something, why not start a discussion with an AI?
Martin Compton has made a useful video demonstrating GAI discussions. You might also like a previous video of his looking at enhancing any lecture notes you might have made.
GAI can help you understand tricky concepts. Hopefully, at this stage of the semester, you have covered the basics, and you do understand them! However, if you are very unsure of some of the content, then a quick question about some of the basics could help you. Remember, GAI can make errors, so perhaps using what it’s generated to “spot the ‘deliberate’ mistakes” could act as a useful tutor for you.
What other ways have you found to use GAI in your revision? Do you think it was useful, or do you have other ways you have found more useful?
Finally, remember, revision is critical; GAI can help you, but it shouldn’t be the only way you revise. You don’t have to use it at all, and if you choose to, it’s most effective alongside a range of strategies.
Some Tips for Using GAI for RevisionThis article was written by Adrian Kakinda (ASC tutor and PhD Candidate in Psychology)
Although GAI is often seen as a writing tool, there are many other ways we can use it at university. GAI can be a useful revision aide, helping to organise our notes, understand complex topics, create practice questions, and even design a revision schedule. While it’s essential to ensure that we don’t use GAI to bypass learning and thinking, there are many ways GAI can make our revision more efficient and free up time for us to use active learning techniques.
In this post, we’ll explore four areas where GAI could make your revision more effective.
The best place to start is with your own notes. Turning your notes from messy lecture scribbles into something usable can be time-consuming. Using GAI, however, you can transform your notes into clear summaries, mindmaps, and flashcards in just a few minutes, leaving you plenty of time to spend on more meaningful, active revision.
GenAI can help structure, refine, and improve your notes by:
Condensing & summarising lecture notes into key takeaways
Rewording complex ideas for clarity
Reformatting messy notes into structured outlines
Highlighting recurring themes across different topics
Transforming notes into flashcards for active recall
Example Prompts:
“Summarize my lecture notes on cognitive psychology in five bullet points.”
“Convert my notes on quantum mechanics into a mind map.”
“Create a glossary of key terms from my notes on international law.”
GAI can also help you learn new information as you prepare for an exam. If you're grappling with a tricky theory or you're not sure if you understand a new idea, GAI can provide explanations or let you test yourself. While it's always important to compare what GAI produces with your notes and course materials, this can be a great way to further your understanding of particularly difficult topics.
GAI can support your learning by:
Testing your understanding: Get GAI to test you by providing it with your explanation of a topic, then asking if it would add or change anything.
Developing original strategies: GAI can create memorisation aides such as mnemonics, musical compositions, rhymes, and anecdotes.
Create learning materials: If you're a visual or audio learner, GAI can design study aides such as diagrams or podcasts that fit with your preferred ways of learning.
Simplifying complex concepts: Stuck on a tricky theory? Ask GAI to break it down as if explaining to a five-year-old.
Explaining in new formats: Whether you prefer mind maps, bullet points, or narrative explanations, GAI can present information in various ways.
Example Prompts:
"My current understanding of shear force is [enter your description]. Is this correct, or have I missed any important points?"
"I'm struggling to understand the theory of relativity. Can you explain it to me like I'm a child?"
"I need to remember the features of melanoma. Can you create a mnemonic to make it easier to memorise?"
Testing your knowledge is crucial for retention. Many courses provide "past papers" to let you practice with real exam questions, but these are often limited and may not be available for all modules. GAI tools can create customised questions from your course content that act much like past papers, allowing you to test your knowledge and practice answering exam-style questions.
How to Use GAI for Practice Questions:
Create Multiple-Choice Questions (MCQs):
Prompt: “Generate five multiple-choice questions with answer explanations on the topic of behavioural psychology.”
Develop Short-Answer Questions:
Prompt: “Create three short-answer questions on the impact of climate change on biodiversity.”
Essay Question Generation:
Prompt: “Suggest two essay questions on artificial intelligence ethics.”
Then, once you’ve written your answers, why not show them to the AI and ask if it would add or change anything? You can then critically evaluate its changes and decide if they’ve improved the answers, or if your original responses were better.
More broadly, GAI can help you create a study schedule to organise your revision. It can often feel impossible to fit everything into the revision period, but GAI can design a schedule that fits within your time constraints. It can’t make more time for you, of course, but it can help you make the best use of the time you have.
|
Example Prompt for a Study Schedule
(edit the bold text to suit your situation): |
I am a [first-year undergraduate] student planning for my exams at a UK university. Create a revision timetable with dates and times. Start the timetable from [Monday 21st April]. Exams start on [Monday 26th May], and there are exams for [Cognitive Psychology 1201, Psychology of Learning 1202, Developmental Psychology 1203]. I have [2] hours available each weekday [avoiding 9am - 5pm] and [4] hours available for each day of the weekend. [Do not schedule any revision between 10am and 4pm on a Saturday due to work commitments]. Focus more time on [Cognitive Psychology 1201 as this is a challenging topic worth 50% of marks for the year]. Plan the schedule to include a maximum of [30] minutes of working followed by a [10] minute break. Present the results in a table format for each week. |
AI-generated schedules can be an effective starting point for organising study plans. It is important to keep your schedule adaptable, just in case something changes in your life, such as a new commitment or an illness. If something does change your plan, though, why not ask GAI to rework it for you?
Using GAI won't make you an expert overnight, but it can greatly improve your study methods. GAI can handle the basic tasks for you, allowing you to devote your attention to the most effective active learning methods. Crucially, though, it can't do the learning for you! Although it can generate practice questions, create flashcards, and help you organise your schedule, it's up to you to actually do the work. GAI won’t make revision easy, but it can be a useful tool to help you get the most out of your study time.
This article was written by Taylor Jeoffroy (PhD Candidate in English, and ASC Tutor)
You’ve probably been warned (more than once) not to use Generative AI (GAI) to write your assignments for you. But did you know that GAI can help you in the pre-writing stages of an assessment? Some of the things GAI can help you with at this stage include brainstorming ideas, finding sources and reading comprehension.
Here are 4 ways in which you can use GAI to help you during the pre-writing stages and some prompts for each to help you get started:
Struggling to narrow down a topic? You can try using GAI to input a broad area of interest and get a list of possible research angles. Using GAI in this way can introduce you to research topics you may not have considered or help you turn vague ideas into more focused research areas. You can ask for themes that are commonly explored in that field or use AI to generate subtopics, such as key debates, case studies, or historical perspectives.
“I’m a first-year university student interested in writing about [subject]. I have the following ideas: x, y & z. Am I on the right tracks? Do you have any further suggestions?”
“I’m a taught master’s student wanting to write a research paper on [topic]. I’ve established x and y as broad themes but I’m not sure how to narrow it down. Can you suggest some specific case studies or debates?”
“I’m a final year undergraduate student on a dissertation module. Give me five unique dissertation topics related to [general area of interest].”
You might find connecting your ideas to established theories and academic frameworks difficult. GAI can help you identify relevant theories or academic perspectives. You can use AI to find key models or frameworks that scholars have used to study it. Now it’s important to note that you still need to put the reading and research in – you’re using GAI here to help get you started and to make connections and fill gaps, not as a substitute for the hard yards of actually conducting the necessary research. You should always critically evaluate AI-generated suggestions and follow-up suggestions yourself, rather than blindly copying them 
“I’m a taught master’s student writing about [research topic]. So far I’ve identified x & y as relevant theoretical frameowrks – would you agree? Have I missed anything obvious - what theoretical frameworks from [relevant discipline] might also apply?”
“I’m a first-year undergraduate student beginning my research on an essay. What are some key theories or academic models that could help get me started analysing [specific issue]?”
“I’m a final year student exploring [general topic]. I’ve read x, y and z. Can you suggest any other relevant theorists, frameworks, or methodologies that scholars commonly use when exploring this topic?”
Finding sources relevant to your research area can be overwhelming, especially if you don’t know where to start. Increasingly, academic databases are adding AI-powered functions that generate search strategy suggestions. GAI can provide a list of foundational texts in a field, suggest keywords, and research strategies. It’s really important to remember though that this should not replace a proper database or library search. Developing effective research skills is a crucial element of any academic degree. So use GAI to supplement these skills, and never as a shortcut or an attempt to by-pass academic databases. And remember, library staff are always on hand to help you develop your search strategies.
“I’m a second-year student researching [topic or question]. What keywords and search terms could I use to find relevant academic sources?”
“What types of sources (books, journal articles, case studies, etc.) are commonly used to study [topic] at taught postgraduate level?”
“I’m a first-year university student. I’m starting to brainstorm ideas for my essay. Can you suggest major scholars or foundational works in the field of [discipline] that I could look up?”
Understanding complex academic texts can be a challenging skill to master. Use GAI to help you check your understanding of an article. Be careful though – you should avoid having GAI generate summaries for you in place of reading the article yourself. Instead, use GAI to help test your own understanding of the article. AI should be used as a thinking partner, not as a content generator
Key Terms and Points:
Read through the article and make a list of key terms and points. Then share your list with GAI and ask it to give you feedback. This can help you check your understanding of the article and see if you are gleaning the main ideas of the article while you read.
“I’m a third-year student studying [discipline]. I just read [article details] on [topic]. Here’s my summary: [insert summary]. Does this capture the key points, or am I missing anything important?”
Generate Questions:
Ask AI to create questions based on an article you’ve read. This can help you check your understanding of an article and may prompt you to think of the article in new ways. This practice can mimic a discussion you might have in class with peers who have read the same article.
“I’m a taught master’s student preparing a paper on [question]. I’ve just read [insert article citation]. Can you ask me three questions about this article connecting to [research topic]? Provide feedback regardless of whether I’m correct or not.
“I’m a first-year [discipline] student. As a follow-up to a lecture on [topic] I’ve just read [insert article citation]. Can you ask me three questions about how this article connects to other areas of this field?”
When we’re just starting out in a field of study, understanding concepts with large bodies of research or technical language can be challenging. Use GAI to help simplify explanations for you to better understand your reading.
“I’m a first-year [discipline] student, and I’m still coming to grips with the subject. This article discusses [concept], but I’m struggling to understand it. Can you explain it in simpler terms?”
