AI-based Group Support to Improve Shared Understanding

A key challenge in group work is not only coming up with an initial topic idea, but developing a shared understanding of what the group wants to achieve. Group members may use the same terms differently, bring different assumptions or background knowledge, prioritize different goals, or disagree implicitly about what would count as a good project outcome. Often, these differences remain hidden until later in the project, when they become harder to resolve. This master project investigates how generative AI can support groups during this clarification phase. 
As part of the thesis, you will design, implement, and evaluate an AI-based prototype that supports group work during this clarification step. The system may, for example, help groups surface implicit assumptions, structure emerging goals, identify unclear concepts, document agreements and open issues. A key goal is to visualize the group’s evolving shared understanding. The system can build on an existing technical infrastructure to listen to in-person meetings. The exact scope of the prototype will be defined during the proposal process.
The project will include:

• Developing and deploying an AI-based prototype to support small groups reach shared understanding as they define and refine a project
• Collecting data during the autumn semester
• Analyzing the collected data
• Writing a final report and giving a presentation

The data collection will take place during the semester. You will need to be present at UZH during October to conduct the data collection.

Timeline: Start as soon as possible

If you would like to work on this thesis, please contact us underhise@ifi.uzh.ch. Please include a current transcript.

The exact scope of the prototype and thesis will be aligned during the proposal process before registration.