Framing AI use in project groups
We are in it together

Framing AI use in project groups
We are in it together

Talking about generative AI
You are sitting with your project group on a Tuesday afternoon, making good progress on your project. One group member suggests using AI to gain an overview of the literature. Another has already used a chatbot to rewrite parts of the initial problem statement. A third is unsure: Are we actually allowed to do that? Should it be mentioned in the project? And what about our shared responsibility at the oral examination? Then someone says: “Perhaps we should ask the chatbot?”
The use of generative AI has become a routine part of many students’ study practices. However, without clear boundaries it can quickly give rise to uncertainty and misunderstandings. It is entirely possible that group members will have different opinions and expectations regarding what they consider fair, academically responsible, and permissible.
Many students at AAU express concern that disagreements regarding AI use can create tensions and conflict within the group. An open discussion in the group may not resolve every dilemma in advance, but it will help you establish common ground. Once you have reached agreement on the extent of your AI use, you can add it to your group contract (cooperation agreement).
Conversation cards to project groups
The purpose of the conversation cards is to reflect on how you want to use AI in your project work, and to simplify making shared guidelines.
How to use the conversation cards
- Create a document entitled: "Our agreement on the use of AI in the group".
- Pick relevant scenarios from your current project phase, or go through all the cards (you can even make your own cards, if you find something missing).
- For each scenario, you write down:
- The group's academic and ethical attitude towards the situation
- Concrete agreements that apply to all group memebers
Most importantly, get started on the conversation about AI use as quickly as possible to avoid uncertainty and misunderstandings in your coming project work.
Scenario 1: Generative AI as the Invisible Group Member
During your conversations and discussions, one group member starts feeding your ideas into a chatbot: “I’ll just ask the chatbot!” It quickly provides definitions, structures and recommendations. The AI-generated responses seem convincing, but the group dynamic begins to change. Other group members stop contributing, and the group's own ideas are not always fully explored.
Questions for discussion:
- When should AI be allowed to participate in the group's discussions and meetings, and when should we allow ourselves to think through the issues first?
- How do we ensure that AI does not take over our academic discussions?
- Would it be useful to introduce a ground rule stating that everyone contributes with their own ideas before prompting an AI tool?
- How do we maintain a balance within the group when some members are highly experienced in prompting, while others have much less experience?
Scenario 2: The boundaries of acceptable and legitimate use
You have just started your project. One group member suggests letting a chatbot produce a draft of the problem statement. Another wants to use it to establish the structure first. As the discussion develops, it becomes clear that you have very different views on what constitutes useful support and when the chatbot responses begin to dilute your own work.
Questions for discussion:
- In which specific parts of the project and project process may AI assist you with e.g. idea generation, structuring, or language feedback?
- Are there core areas of the project where we deliberately choose not to use AI in order to support our own deeper learning?
- Where is the boundary between receiving support and allowing the model to take over the task?
Scenario 3: Being open and trusting each other's contributions
Your group meets to compile a section of the project’s theoretical chapter. One group member brings along a highly polished and well-written piece of text, mentioning in passing that AI “helped a bit along the way”. Some group members become uncertain about how much of the text was actually written by the individual and whether everyone is contributing on equal terms?
Questions for discussion:
- What does it mean in practice to be transparent with one another about the use of AI?
- When does AI use become something that should be disclosed to the rest of the group?
- How can we foster an open culture in which it is acceptable to share how AI has been used?
- Dare we challenge a group member who presents a highly convincing AI-generated output if we have doubts about its academic accuracy?
Scenario 4: Fairness and shared responsibility in AI-supported group work
The group divides the workload: two members spend days reading complex articles on the philosophy of science and its connection to the project's methodology chapter, while a third member quickly completes her part using automated summaries and AI-generated text.
Questions for discussion:
- How can we ensure that AI does not create an imbalance in our individual contributions and workload?
- Do we agree that AI-powered shortcuts should not replace thorough engagement with and reading of mandatory course literature?
Scenario 5: Shared responsibility and exam
The project includes an analytical model whose structure was generated with the assistance of AI. During the oral examination, one group member is asked about the methodological choices underpinning the model and struggles to explain and justify them.
Questions for discussion:
- How do we make sure that every member of the group understands and is able to explain and defend all sections of the project during an oral examination?
- What steps should we take to verify and validate AI-generated content before relying on it in our final report?
Scenario 6: Declaring the use of generative AI
The submission deadline is approaching, and you realise that members of the group have different understandings of what aspects of AI use should be declared and how this should be done in practice. As the project has been underway for some time, it is difficult to remember which tools were used and exactly how they contributed to the work.
Questions for discussion:
- Do we know the formal requirements for declaring the use of AI that apply to our module and study programme?
- How do we establish practical routines to make sure that everyone documents their use of AI from the outset (for example, by maintaining a shared prompt log)?
Scenario 7: Climate impact and active rejection of AI
Your group is about to begin an extensive text analysis. Two members suggest uploading the texts to an AI service to quickly generate summaries, tables, and academic comparisons. A third group member objects, pointing out that large language models have a significant carbon footprint. He proposes, on ethical and environmental grounds, that the group should refrain from using AI for this task and instead divide the texts among the group members, read them independently, and compile a shared document. The group is divided between considerations of efficiency and concerns about climate impact.
Questions for discussion:
- To what extent do environmental and climate considerations influence our decisions to use, or not use, AI in our studies?
- What criteria should we use as a group to decide when the use of AI is warranted, given both its academic value and its environmental impact?