Norwegian version of this page

How to Use AI as a Student

On this page, you will find examples and advice on how you as a student can use AI, in ways that promote learning.

a group of students in a dialogue with a teacher
Photo: Jarli &Jordan/UiO
GPT UiO is one of the approved AI-tools at the University of Oslo. It is an UiO-developed tool that allows you to use OpenAI's ChatGPT within the privacy requirements set by UiO. The purpose of this website is to provide you with advice and examples on the use of this AI tool, which is built on language models. It's important that you check with your teachers what they consider acceptable AI use in their courses.

Learning with GPT UiO

GPT UIO is a useful tool that can contribute to your learning outcomes. However, if you let the tool do the work for you, you could miss out on a lot of valuable learning. It is you as a student who should manage your own knowledge development. 

 Excessive use of AI-generated texts can hinder your development of  analytical skills and critical- and creative thinking. Writing and editing one?s own text is often important for developing subject understanding and one?s own thinking in additiond to your academic writing skills. The purpose of writing assignments at the university is not primarily to assess what you have learned, but to give you opportunities to develop your skills in academic writing and thinking. It is incredibly important that if using AI, you put yourself in the driver's seat of the writing and use AI as a tool and a support to become better at writing.  

 GPT UiO can create summaries and simplifications of large and difficult texts. This can be useful. But it is also very important to read the texts yourself and process the content. This is how you get a good understanding of academic work in your course and can evaluate the content and what it means. When working with text-generated content, it may be wise to keep these questions in mind

  • Can I trust the answers GPT UiO gives?
  • Do I have the opportunity to check the content against original texts?
  • Should I ask the question in a different way to get a more extended or specific answer?

Understand the Technology

It is particularly important to understand that the technology behind some AI tools, such as GPT, is based on language models. This means that they are not knowledge-based, but the answers they provide are based on statistical forecasts that the model is trained on. The language model  

  • looks at patterns in the words you use in your questions and calculates what the model thinks are the most important words in the sentence to give an answer that is likely to make sense. For example, the model can calculate whether the word "bark" in the term "the bark was old and gnarly" means the protective outer covering of a tree or the sound a dog makes.  
  • compares this calculation with the statistical pattern the model has from training.  
  • calculates the probability that certain words may be appropriate in an answer and gives you the most likely words back.    

Remember that if some words are likely, does not mean that the answer is true. If you ask a language model to complete the sentence "The Earth is...", and the language model is trained on many texts containing the sentence "The Earth is round", it will suggest the word "round". If it is trained on texts that say the Earth is flat, it will suggest that instead.  

No Definite Answers

In other words, there is no form of fact-checking involved when a language model like GPT UiO responds to your questions, just a probability calculation based on texts it has analyzed. We do not know which texts GPT UiO has been trained on, neither do we know exactly how it has been trained. The same applies to most other language models. Therefore, we cannot rely on the information they provide us as true.

    Basics and Useful Resources

    Here are some entrances to a basic understanding of the technology. Once you have understood the basics, the next step is to understand and experience how available tools can be used in your course. See our selection of resources that can help

    1. The video Introduction to AI for Teachers and Students. Produced by Wharton School. The video explains what artificial intelligence is. The video illustrates other AI tools than GPT UiO, but is based on the same language technology. Viewing time is 10 minutes. Available in English. 
    2. The website AI ethics and Social Responsibility presents the critical aspects that you as a student at UiO need to consider. Estimated reading time is 10 minutes. Available in Norwegian and English. 
    3. The video Prompting AI. Produced by Wharton School. The video shows how you can ask questions, read prompt, in a useful way. The video shows other AI tools than GPT UiO, but you can use the same approach. Viewing time is 11 minutes. Available in English. 

    Learn more

    • The video Large Language Models. Produced by Wharton School. The video gives you an explanation of large language models. The video illustrates other AI tools than GPT UiO, but is based on the same language technology. Viewing time is 12 minutes. Available in English.   

    • The report ChatGPT and Artificial Intelligence in higher education: Quick start guide issued by UNESCO gives you a short description of what artificial intelligence is, examples of use, and outlines some critical reflections one should be aware of. Estimated reading time is 30 minutes. Available in English. 

    • The website What are AI tools and how do they work? created by University of Groningen gives you a short explanation on how AI tools function, and which AI models exist. Estimated reading time is 5 minutes. Available in English.  

    Explore and Use

    Before you start, ask teachers at your course what they have decided is acceptable use of AI in the course. Then, explore with fellow students how this technology can be used to support your work. Here are some questions that might be relevant to work with in such a process  

    • How can the tools support learning in your course?   

    • What are the tools good at? What are they bad at?   

    • How can using the tools contribute to, and elevate your learning?    

    • What is important so that the use of the tools does not make you learn less?   

    • What ethical concerns are relevant?   

    • What does ethical concerns mean for how you use the tools?  

    Prompting

    When speaking with language models, such as GPT UiO, it is important to be able to ask specific questions that contain a lot of context. This action is called prompting. In Norwegian, prompting can be explained as question formulations, instructions, or cues. Prompting is your way of giving GPT UiO instructions on what you want it to do.  

    It can be challenging to create prompts that provide you with the desired response. GPT UiO bases much of its answers  on the information and questions you provide. Therefore, there is often need for adjustments,corrections, and follow-up questions in your prompts The stronger your grasp on the subject matter, the higher the likelihood of producing useful prompts. The most important thing is not to ask the perfect questions, but to try out and test. In some cases, you might get more accurate results if you prompt in English. For such AI tools to be useful to you, it is important to view the interaction as a dialogue. If you want to learn more about this, go to the "Good to know section" on this webpage.  

    Proper, critical and honest use

    Make sure you are well acquainted with the guidelines in your course. Ask if you are unsure. In general, this applies to all work that is assessed. For example, exams should be your own work, even though you are often allowed to use exam support materials. It is teachers at your course who decides if and how you might use AI.  

    To use AI-generated text in an academically responsible manner in your writing process, you must actively adhere to academic norms for honesty, accuracy, and transparency. This means that it is important that you

    • clearly indicate when and how you use such texts  

    • critically assess and verify information from AI sources   

    • do not rely on AI as the only or primary source of information  

    If the task consists of writing a text, this will, for example, entail that you show in your text which sources you are building upon. You must further clarify for the reader the difference between your own voice and the voices of the authors you are referring to. You can read more in search and write resource on how you can practice clearly expressing your own opinions and what you are building upon from others. 

    •  If you use text generated by AI, for example GPT UiO, in the process of developing your own text, you must be aware that such AI-generated texts do not adhere to academic norms for accuracy and transparency.   
    • AI-generated text may contain errors, inaccuracies or be misleading. So always verify the text with several other sources.  
    • AI-generatied text t is not your own. If you use it in your work, you need to be open about which parts of the text is AI-generated, how it was generated and used in your work.    
    • AI-generated text usually does not refer to sources, or sources it refers to are not necessarily real or relevant. . To be an honest writer, you need to find, explore and reference real academic sources.  
    • AI-generated text may reflect biases or prejudices from the training data. By building on biases or prejudices ,you may contribute to   reinforce such bias and prejudices. .   
    • Do not use AI to write the text for you, but as a support in the writing process, for example to get ideas and improve your own text.

    If you use AI in the process of developing your own text, you should be open about this, and clearly show how you have used AI and what is your independent contribution in the development of the text. Here you can read more about how you can be transparent.

    In addition to the resources available through the University Library's Search and Write, UiO has a shared website about use of source and references. It is also important for you to familiarise yourself with the rules regarding the use of aids during exams, to ensure there is no suspicion of cheating or attempted cheating.

      Examples of Using AI as learning support

      For useful answers from language models like GPT UiO, it's better to be as concrete as possible in your inquiries and include context for more precise responses.Additionally, it is important that you do not include sensitive personal data, such as details about an individual's health, religion, political views, or similar. Read more about this and other legal guidelines for the use of AI at UiO. Before going through the examples below, remember that GPT UiO does not have all the answers, and like with all other sources, you need to critically assess the content it produces. In addition, it's important that you ask your teachers about what is considered responsible AI use in your course.  

      Writing Support

      GPT UiO can provide feedback on written text or ideas for your text. GPT and similar tools are good with language, especially English, but are more varied on content. Therefore, you should critically assess the feedback you receive and consider how you process your texts. Here are our suggestions for use

      • Ask for suggestions on how your text can be improved. Ask for specific improvements, such as clarity, making the text shorter, more varied language use, good transitions between paragraphs, or the like.   

      • Ask for suggestions for alternative ways to express an argument or idea.   

      • Ask GPT UiO to oppose the arguments in your text, in your process of developing and nuancing your own argumentation.  

      To know whether the suggestion is good and correct, you must have insight into the topic. If you do not find the text that is generated to be relevant for what you want to write about, you can provide more specific prompts. In this way, you get to reflect on what you want to write. Never use the generated text directly but pick out and rewrite the parts that help you progress with your own writing. 

      Tip!

      Try multiple times. Take out what you find helpful to express what you want in your text and get a form and tone you experience as your own. 

      Evaluate your sources

      It is important that you use and refer to good sources if the content/arguments in your text change during the process. 

      We recommend this article if you want to read more about “how to use AI as writing support”: Anson, C. M., & Straume, I. (2022). Amazement and Trepidation: Implications of AI-Based Natural Language Production for the Teaching of Writing. Journal of Academic Writing, 12(1), 1–9.

      Study Group Partner

      You can include GPT UiO as a colloquium partner to get different input and perspectives that you can discuss and evaluate. Having academic discussions together with GPT, and evaluating the input GPT generates, can help lift the discussion and learning in your colloquium group. Here are our suggestions for use

      • You can discuss academic terms, claims, historical timeline etc.  

      • You can ask GPT UiO to create case assignments and reflection assignments that you can discuss and reflect on together.

      Tip!

      Be aware that GPT is trained to “please” you, and you cannot blindly trust the content. It is therefore a strength to discuss the content you receive with your study group.

      Coach

      You can use GPT UiO as your own coach in your learning work. You can get advice, input, explanations, have discussions, or get step-by-step guidance. Here are some suggestions for use

      • You can enter into GPT UiO your own understanding of the phenomena you are studying, or how far you have come with a task, and ask for suggestions on how to proceed further.    

      • You can make GPT UiO tailor content to your learning level. For example, "I am a beginner in course x, can you explain to me what the term x means?", or "Can you simplify this text so that a beginner can understand?".   

      Tip!

      Evaluate the advice you get and choose the ones you find good. Feel free to ask for better, more specific advice. And feel free to discuss the advice with fellow students!

      Programming

      You can use GPT UiO to learn programming and coding. It is about understanding concepts, the ability to break down problems into smaller parts, and practicing the ability to find and understand errors in the code. It is also about getting the programming language under your skin.   

       Clarify with the teacher what is considered acceptable for use in your course. Can you submit code written by GPT? What if the comments in the code are written by GPT? Is it okay to use GPT to break down the task into manageable tasks, or in other ways use GPT to understand the task? Here are our suggestions for use  

      • If you have written code that works, ask GPT UiO to give you feedback on the code and suggest changes for improvement.    
      • Get feedback on a code that doesn't work as expected or gives an error message, and you cannot understand why this is happening.  
      • If you have received a task, and even after you have thought thoroughly about the problem cannot see how the task can be broken down into manageable pieces and you do not know where to start, then a way forward could be to ask GPT UiO for tips. Feel free to ask GPT UiO for tips on how you can manage to do it and ask it to create a step-by-step explanation.  

      Examples of prompting 

      • I am learning programming. Can you explain recursion?
      • Can you explain it another way?
      • How does the computer handle recursion?
      • What are reasons to use recursion, and what are reasons to not use it?

      Tips!

      Quantity training is important, so do not try to get GPT UiO to do all the work for you. You want learn by yourself. And even though GPT UiO will be part of a programmer's toolbox, it is important to check that what GPT does is correct and good. In addition, it is important to be able to guide GPT towards the result you want, and to be able to write or change code if you don't get the desired result.