Reflective Writing for Personal Development.
An innovative conversational agent and data dashboard, based on cutting-edge Gen-AI technology, will be developed to support learners' skill development through personalized reflective writing.
Factsheet
- Lead school Business School
- Institute(s) Institute for Digital Technology Management
- Strategic thematic field Thematic field "Humane Digital Transformation"
- Funding organisation Innosuisse
- Duration (planned) 01.08.2024 - 31.01.2026
- Project management Prof. Dr. Roman Rietsche
- Head of project Prof. Dr. Roman Rietsche
-
Project staff
Prof. Dr. Roman Rietsche
Léane Marie Klara Wettstein
Situation
A workforce with metacognitive skills is critical to achieving the United Nations Sustainable Development Goals (SDGs) and adapting to changing skills in a changing economy. The World Economic Forum estimates that 71% of workers will require significant upskilling. Current education systems struggle to meet this need due to high costs and limited scalability. Reflective writing is critical to building metacognitive and sustainable skills, but universities often lack the resources for the personalized feedback needed for effective training. To overcome these challenges and support the development of reflective writing skills, we are introducing two innovations to the Rflect AG app: a conversational agent, which provides tailored, step-by-step feedback, and a journey dashboard, which provides students with visual insights into their progress and faculty with grade-level tracking of their skills. These tools make reflective writing more accessible and effective for both learners and institutions.
Course of action
First, an interactive dialog system is developed to deepen the learners' reflection experiences. The conversational agent assesses the depth of reflections and adapts to learners' reflective abilities to provide them with helpful feedback. The system is based on the Gibbs reflection cycle and uses advanced prompt engineering techniques to encourage deep reflection. Different prompt engineering strategies will be tested and iteratively improved to increase the adaptability of the system. An initial version will be integrated into the Rflect app in the fall semester, and enhanced versions will be released over the following two semesters based on feedback and performance data. In a second step, a dashboard will be developed. The journey dashboard will give learners insights into their reflective skills and sustainability competencies and provide teachers with an overview of progress. The dashboard uses AutoGen to analyze reflection texts and identifies patterns and themes in unstructured data. Large language models (LLMs) are used to perform thematic and sentiment analysis to monitor the development of reflective competencies. An iterative process is also used to test different versions over several semesters.
