On September 8, the workshop on Learning Analytics 2025 is planned as a half-day workshop taking place during the DELFI’25 in Freiberg. The topic of this year’s workshop is »Learning analytics in the era of large language models«. In addition, other relevant topics on learning analytics are welcome. We are currently accepting submissions to the workshop until Juni 15. Please refer to the call for papers for further details.
Workshop topic
A central aim of learning analytics is to adapt learning and instruction to learners’ needs. However, learning analytics in the past have often only provided basic visualizations representing users’ online behavior instead of providing detailed feedback. Language models have been investigated for a long time. Since 2022, with the rise of GPT-3 (Generative Pre-training Transformer 3) language model, the term of Large langue models (LLMs) started to appear. With the advent of ChatGPT the affordances of LLMs are also increasingly discussed in the broader educational context. Generative AI might be one solution to optimize the personalization at scale that learning analytics aims to offer learners. This could include personalized recommendations and feedback but also the system could explain its underlying mechanisms in different complexity to its users. Furthermore, besides only providing feedback on indicators that are easy to measure, learning analytics could, with the capabilities of LLMs, also analyze unstructured data and provide feedback.
Analyzing learners’ prompts might offer valuable insight into their needs and learning processes. On the other hand, learners can also use generative AI to solve their tasks, which may result in biases in the performance data in the underlying learner model, emphasizing the relevance of digital learning behavior within the digital learning environment.
The idea of how learning analytics and generative AI complement each other is a current topic in the field but has just recently begun. Hence, there are a number of questions about how learning analytics in the era of generative AI might evolve and how the two areas can enhance each other: How can generative AI support the processes of grading and feedback in learning analytics systems? How can generative AI support the understanding of learning analytics results by the stakeholders? How can generative AI support learners in using the results to adapt their learning processes? How can generative AI support educators in adjusting their instruction to learners’ needs? How do learners interact with generative AI and what measures could be added to learning analytics models to gain insights into learners’ difficulties and learning paths? How can generative AI bias learning analytics? How can generative AI support data collection and analysis?
Goals of the workshop
The overall goals of this interdisciplinary workshop are:
- Networking of the Learning Analytics community in the German-speaking countries to initiate research and joint projects as well as fostering network activities beyond this workshop;
- Presentation of current research projects and results;
- Enhancing the interdisciplinarity in the working group: e.g., computer sciences, artificial intelligence, math, learning sciences and psychology as well as ethics and philosophy;
- Developing a joint outcome (e.g., paper, digital learning resources, event) based on the workshop discussions.
Programm Committee of the Workshop
tbd
Organizers
Clara Schumacher (University of Potsdam)
Jakub Kuzilek (Humboldt-Universität zu Berlin)
Claudia Ruhland (Humboldt-Universität zu Berlin)
Leo Sylvio Rüdian (Humboldt-Universität zu Berlin)