This project sought to develop and evaluate an AI system designed to support and enhance the most important feature of the collegiate University: interdisciplinary discussion.
In an era when students are increasingly encouraged to interact with AI in isolation, the ‘AI for Interdisciplinary Discussion’ project set out to do something different - using AI not to replace human conversation, but to spark and support it. The pilot platform was designed to analyse examples of written work uploaded by users and generate tailored prompts and questions to fuel face-to-face dialogue among postgraduate students. The project aimed to encourage the kind of interdisciplinary debate, social connection, and peer learning that are vital to the University experience.
With technical development led by Haseeb Ahmad, Senior Research Software Engineer at the AI Competency Centre, the team built a bespoke AI platform and interface. Off-the-shelf AI tools were unsuitable for the project’s needs, as the functionality required demanded a custom design.
Two rounds of student feedback shaped the pilot. The first involved an in-class demonstration by four members of the project team, followed by a full-group discussion. Students responded with strong interest, particularly to the idea of AI as a facilitator for social connection and collaborative learning. Many expressed enthusiasm for expanding the tool into a broader community space for interdisciplinary exchange.
Key lessons emerged from the process. The unconventional nature of the project meant there were few existing AI solutions to adapt, making bespoke development necessary - but also more time- and labour-intensive. The project demonstrated that, with the right design, AI can help foster human-to-human dialogue rather than diminish it, offering a model for tools that connect students across disciplines and encourage richer, more inclusive academic conversations.
Exploring AI-Human Triangulation for Research Reproducibility
AI in Education