So What? Unpacking the Complexities in Collaborative Problem Solving with AI-Augmented Sense-Making

Abstract

Visual representations of data in dashboards have the potential to translate learning process data into actionable insights for educational stakeholders. This is particularly useful for teachers when their capacity to monitor students is further strained in orchestrating complex activities, such as paired collaborative problem-solving. Despite great interest in dashboards over the past decade, their effectiveness has been under scrutiny with concerns about the cognitive overload for users with limited data literacy, thus questioning their practical utility in supporting decision-making and reflective practices. In this paper, we present a functional prototype for K-12 teachers and demonstrate the use of generative AI to unpack rich details of collaborative problem-solving processes in mathematics. Through a case study, we share our human-centered approach to design, ensuring that AI-augmented insights are not only interpretable and actionable, but centered around authentic needs. Based on the insights and experiences in the co-design session, our tool includes features, such as a contextualized description of students’ breakthrough and struggle moments. This work contributes to ongoing efforts to support users’ sense-making process of data visualizations in dashboards by leveraging generative AI.

Materials

• Full BibTeX (publisher)citation-needed.springer.com

Cite

@inproceedings{lee2025so,
  author = {Lee, Seiyon M. and Li, Hongming and Zhang, Shan and Zhong, Zirui and Lee, Ji-Eun and Botelho, Anthony F.},
  title = {So What? Unpacking the Complexities in Collaborative Problem Solving with AI-Augmented Sense-Making},
  booktitle = {Artificial Intelligence in Education},
  year = {2025},
  volume = {15881},
  pages = {431--438},
  publisher = {Springer Nature Switzerland},
  doi = {10.1007/978-3-031-98462-4_54},
  url = {https://link.springer.com/chapter/10.1007/978-3-031-98462-4_54}
}