Make it Make Sense: An Exploration of AI-Mediated Individual Sensemaking in Education Through the Lens of Cognitive Mechanisms
Tiu, Carlos (2025)
2025
Master's Programme in Computing Sciences and Electrical Engineering
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
Hyväksymispäivämäärä
2025-12-17
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-2025121711848
https://urn.fi/URN:NBN:fi:tuni-2025121711848
Tiivistelmä
Sensemaking is defined as the way individuals interpret, connect, and relate new infor-mation around themselves with their existing knowledge and past experiences to build up their understanding of reality. Because it is fundamentally connected to learning, it is an important lens with which to understand a student’s studying and analysis process. Zhang and Soergel’s model of individual sensemaking was used as a foundation for understanding sensemaking in this study because it outlines sensemaking activities in general through the cognitive mecha-nisms employed to instantiate, tune, and adjust an individual’s knowledge structure.
In light of the rise of AI use by both students and educators for a myriad of purposes, this study aims to observe AI-mediated sensemaking for the specific task of academic reading, a common task required of the average student, especially in higher education. By doing so, the study also seeks to observe whether Zhang and Soergel’s model still appropriately captures sensemaking when AI-mediation is involved, or whether there are now changes in the type, frequency, and order of activities that were not apparent before.
A contextual inquiry and interview was conducted on 5 masters students in Tampere Uni-versity’s Human-Technology Interaction program to understand the sensemaking activities they would perform with and without AI mediation, such as the use of LLMs and text overlays such as Semantic Reader. Screen recordings of the tasks, transcripts of think-aloud and inter-view dialogue, as well as note-taking artifacts of all participants were collected and analyzed. Results showed that while there appeared to be no new mechanisms being used, participants tended to offload certain lower-level mechanisms such as summarization, key item extraction, categorization, and restatement. The users’ overall sensemaking loop also appeared to put a greater emphasis on establishing a clear frame of mind to compare and evaluate AI responses with their own understanding of the information being collected.
The work illustrates not only a snapshot of current use cases and capabilities for AI but al-so the implicit and explicitly stated preferences of the users in higher education towards AI tools. While AI can already be used as a tool that offers more structured feedback and oppor-tunities to extend learning, other factors such as the nature of the task given to participants, time pressure, interest in the papers being read, and prior knowledge of the content also in-formed the level of effort and focus participants felt was necessary to dedicate to the tasks. Further research can be pursued towards looking at AI-mediated group sensemaking, a com-mon occurrence especially for learning in classroom or lecture-based environments.
In light of the rise of AI use by both students and educators for a myriad of purposes, this study aims to observe AI-mediated sensemaking for the specific task of academic reading, a common task required of the average student, especially in higher education. By doing so, the study also seeks to observe whether Zhang and Soergel’s model still appropriately captures sensemaking when AI-mediation is involved, or whether there are now changes in the type, frequency, and order of activities that were not apparent before.
A contextual inquiry and interview was conducted on 5 masters students in Tampere Uni-versity’s Human-Technology Interaction program to understand the sensemaking activities they would perform with and without AI mediation, such as the use of LLMs and text overlays such as Semantic Reader. Screen recordings of the tasks, transcripts of think-aloud and inter-view dialogue, as well as note-taking artifacts of all participants were collected and analyzed. Results showed that while there appeared to be no new mechanisms being used, participants tended to offload certain lower-level mechanisms such as summarization, key item extraction, categorization, and restatement. The users’ overall sensemaking loop also appeared to put a greater emphasis on establishing a clear frame of mind to compare and evaluate AI responses with their own understanding of the information being collected.
The work illustrates not only a snapshot of current use cases and capabilities for AI but al-so the implicit and explicitly stated preferences of the users in higher education towards AI tools. While AI can already be used as a tool that offers more structured feedback and oppor-tunities to extend learning, other factors such as the nature of the task given to participants, time pressure, interest in the papers being read, and prior knowledge of the content also in-formed the level of effort and focus participants felt was necessary to dedicate to the tasks. Further research can be pursued towards looking at AI-mediated group sensemaking, a com-mon occurrence especially for learning in classroom or lecture-based environments.