LAK Workshop
Dublin, Ireland | March 3, 2025
Description
As Multimodal Learning Analytics (MMLA) evolved, it became crucial to critically assess its methodologies, especially as Generative AI (GenAI) transformed educational data collection and analysis. GenAI offered significant potential for processing diverse multimodal data, enabling richer insights and personalised learning experiences. However, this rapid integration also brought challenges, including ethical concerns, algorithmic bias, transparency issues, and privacy concerns. Addressing these complexities was essential to ensure that GenAI enhanced learning without perpetuating inequities.
The workshop explored how GenAI could reshape MMLA research, addressing both opportunities and challenges. Researchers and practitioners discussed practical approaches to responsibly integrating GenAI into MMLA, focusing on ethical, scalable, and transparent AI-powered tools. This workshop was built on the CROSSMMLA series’ legacy, offering insights into how GenAI enhanced learning analytics while maintaining ethical standards and promoting equitable access in education.
The half-day workshop accommodated up to 40 participants, including newcomers and experts in multimodal analysis. Participants did not need technical expertise but were expected to be interested in methodology and research design in MMLA. The workshop included reflective and hands-on activities to develop a deep understanding of GenAI research in MMLA.
Workshop Objectives
The workshop equipped participants with practical skills for integrating GenAI into MMLA research. Participants developed hands-on expertise in automating data collection, analysis, and feedback generation across multiple modalities while addressing bias, privacy, and scalability challenges. Ethical considerations were a significant focus, including fairness, transparency, and societal impact.
The workshop also aimed to generate best practices and methodological insights for implementing GenAI in educational contexts. Discussions fostered multidisciplinary collaboration and contributed to the learning analytics community through a publication or report summarizing key insights, challenges, and solutions.
Details of Activities
Elevator Pitches and Framing Discussions (30 minutes): Participants introduced themselves and shared brief pitches of their research interests. Facilitators provided key definitions and curated research examples to frame discussions and encourage active participation.
Diving into GenAI Research in MMLA (1.5 hours): Organizers presented examples of GenAI tools applied in MMLA research, followed by hands-on activities where participants could apply these tools to their own or provided datasets.
Future Considerations for GenAI Research (1 hour): Participants broke into small groups to discuss challenges and opportunities for applying GenAI to MMLA, focusing on themes such as data privacy, algorithmic bias, environmental impact, and transparency.
Reflections and Next Steps (30 minutes): The workshop concluded with a reflection survey and final discussion that summarized the outputs and insights gained.
Organizers
Namrata Srivastava – Vanderbilt University & Monash University Roberto Martinez-Maldonado – Monash University Daniele Di Mitri – DIPF | Leibniz Institute for Research and Information in Education Daniel Spikol – University of Copenhagen Vanessa Echeverria – Monash University & Escuela Superior Politécnica del Litoral Kateryna Zabolotna – University of Oulu Simon Knight – University of Technology Sydney Mohammad Khalil – University of Bergen Gloria Fernandez Nieto – Monash University Ruth Cobos