Artificial intelligence has developed rapidly in recent years, especially in the development of applications based on Large Language Models that enable Generative Artificial Intelligence (GAI).
GAI technologies, such as chatbots, are characterised by their ability to generate text, images, synthetic data, code or other media using generative models via a user-friendly interface. For example, GAI technologies can write new text in response to the questions (prompts) you ask, and they can also edit and rewrite existing text. GAI is able to do this because it has "learnt" patterns and structures from existing text examples and can generate new coherent texts from them.
GAI can be a valuable technology in your studies if you use it carefully, thoughtfully and critically. For example, the technologies can support your workflow to generate ideas, provide creative input and improve your texts.
How to use GAI in your individuel work?` | |
|---|---|
Feedback on your texts
| You can give the chatbot an excerpt from a text you have written, for example from a report or an analysis, and ask the chatbot to give you feedback on the grammar, coherence and clarity of the text. In other words, you can use chatbots to help you improve the quality of your writing. |
Alternative wording
| You can ask chatbots to suggest different ways of wording parts of your text, for example the title or headings, or to phrase your research question in different ways. |
Get over writer's block
| Chatbots can help you get started on a new writing assignment by giving you initial drafts of sentences or passages that you can use to get started. This can be a good tactic if you’re struggling with writer’s block. |
Understand a topic better
| Chatbots can be a good tool to help you get a quick overview of a complicated topic, theory or concept before you investigate it in more detail and depth. But be aware that the information it gives you may not necessarily be correct. |
Improve your reading process
| Before you start reading a complex primary text, you can use chatbots to get a basic grip on the main ideas, in the same way you use internet searches. This can make it easier to understand the text and clarify what you need to get out of it and how to approach it. |
Ude chatbots as a topic
| Chatbots are a current topic of interest, and it might be relevant for you to analyse them in a paper or use theories from one of your courses to discuss them. Interesting perspective include law, philosophy, ethics and pedagogy. |
Identify holes in your knowledge
| Ask the chatbot to explain some academic material to you, and assess whether you – or the chatbot – lack knowledge about it. Maybe the chatbot will say something inaccurate, or maybe it will provide a different perspective on the material. |
Start your search
| You can use chatbots to get an introduction to a topic and get inspiration for specific keywords you can use when searching the literature. |
Programming tasks
| GAI can generate examples of codes or even entire functions based on natural language descriptions or specific requirements. Autocompletion tools can predict and suggest examples of code, making it faster to write code, and code review tools can analyse code for potential errors and vulnerabilities. Finally, GAI can help generate and update documentation by analysing code and identifying relevant information. |
Data analysis
| GAI can analyse historical data to predict potential problems, estimate project deadlines and optimise resource allocation. |
Create diagrams
| Unified Modelling Language (UML) diagrams can be created by GAI using description language and more general illustrations (images) can be generated by AI image generators. |
How to use GAI in your study group? | |
|---|---|
Discuss
| Discuss how you can use chatbots responsibly as a resource in your work together without the risk of cheating. Draw on each other for inspiration – and ask your lecturer for help if you have any doubts. |
Mythbuster
| Confirm or reject the chatbot’s answer. Ask the chatbot a number of questions about a subject you’re studying, and assess its responses critically. Confirm or reject the chatbot’s answer based on your reading of relevant academic sources. |
The turing test
| The Turing test is a way of determining whether a machine can exhibit intelligence equivalent to a human Come up with a series of questions about your subject, as if you were designing an exam. Distribute the questions among yourselves. Answer your question, and then ask the chatbot the same question. You and your fellow students should take turns reading your own and the chatbot’s answers to each other. The other students should try to guess whether each answer was generated by the chatbot or written by the student. |
Topical
| Chatbots are a current topic of interest, and it might be relevant for you to analyse them in a paper or use theories from one of your courses to discuss them. Interesting perspective include law, philosophy, ethics and pedagogy. |
"I've used GitHub Copilot for code generation in different programming languages. I mainly use it to generate code for simple tasks and things I've programmed before. It also helps me to improve code and resolve error codes. The tool has saved me time and it streamlines my programming process." - Bachelor’s student in Computer Technology (4th semester) |
"I used the bot to find a summary of John Dewey's educational philosophy, as I couldn't process the massive number of search results that appeared on Google. So I used ChatGPT to find the ‘canned’ version, and then I could easily continue my search using the different buzz words on Google." - Master’s degree student, IT and Designs for Learning (2nd semester) |
"I used the chatbot to clarify concepts and theories and to set up a discussion about the two theories I used in my assignments to see whether I was on the right track in my own discussion. This confirmed that my understanding of the theories and their differences and similarities was legitimate, but it did not give me a complete answer. So I've used it to develop my academic understanding while writing the assignment." - Bachelor’s student in Educational Science (6th semester) |
"I've used ChatGPT to revisit and get an overview of Max Weber and his theories instead of going back to the heavy text again. The results were of varying quality. Generally in terms of theories, GPT can point you in the right direction, a bit like Google Scholar, but I've also experienced it can be wrong." - Master of Arts student (2nd semester) |
It's important to be aware of both ethical and environmental challenges when using generative AI in your studies. AI technologies can impact society in many ways, including privacy, accountability and environmental impact. This list is an overview of the key points to consider when integrating with AI:
Ethical and environmental aspects of using GAI | |
|---|---|
Enviromental impact
| Several environmental implications stem from the large amounts of energy and resources required to train and run AI models. Therefore, consider whether it is necessary to use GAI and use conventional search engines if possible. |
Bias and discrimination
| Generative AI systems can inadvertently carry forward any bias in their training data, and this can result in discriminatory results. Bias is not immediately visible, but is embedded in responses and decisions. |
Accountability and transparency
| Generative AI is often a “black box”, making it difficult to understand its decision-making processes, especially in critical areas such as medical diagnoses and administrative decisions where AI recommendations can have serious implications. |
Manipulation and misinformation
| AI can be used to create and spread false information such as deep fake or algorithmically generated fake news. This can influence public opinion and social debate. |
Exploitation and concentration of power
| Large tech companies dominate AI development, which can have a major effect on the market, inhibiting competition and reducing innovation in the wider AI landscape. |
Spam flooding
| Generative AI can create huge amounts of content, including spam, which can strain digital platforms and damage information quality. |
Privacy and surveillance
| AI's ability to analyse large amounts of data entails risks of privacy violations and unauthorised surveillance, especially on social media. |
Copyright challenges
| Generative AI models may be trained on copyrighted materials such as books, journals, images, films, and music. In the absence of regulation and transparency in this area, there have unfortunately been instances of AI models violating copyright laws. It is therefore important to be aware that some generative AI models do not comply with copyright regulations. |
The list above was compiled with inspiration from the TeachGenAI course, which students can access via Brightspace.
More environmentally friendly use of GAI | |
|---|---|
Use GAI wisely
| Consider whether using GAI is absolutely necessary for your task or project. Avoid unnecessary use and be aware of when you really need AI tools or if your query can be replaced with a search engine. |
Create good prompts
| When using GAI, try to be precise and target your queries to avoid unnecessary repetition and wasted resources. Practice your prompting skills on the TeachGenAI course |
Share with your group
| If you are working in groups, try to share AI usage where possible. Instead of multiple students making the same queries, one person can use AI and share the results with the rest of the group. |
From the autumn 2024 semester, students are allowed to use generative artificial intelligence (GAI) in all exams at AU unless it is explicitly stated in the academic regulations or course description that they may not. The new rules have been implemented in individual academic regulations and course descriptions, so it may be a good idea to check the course catalogue to find out what applies to your course.
When using GAI, it is extremely important to familiarise yourself thoroughly with the academic regulations and the elaboration in the course description of the exam activities and not least with AU's guidelines for GAI.
The rules for using GAI in exams may change. That's why it's essential to keep yourself updated here:
When using GAI in the exam, you should be aware of the following: |
|
