TLDR: A study by Huang and Willems explores how educators designed custom GPTs for a Master’s-level Qualitative Research Methods course. These AI tools helped students with research questions, interview practice, and data analysis, enhancing reflexivity and analytical thinking. While beneficial for personalized feedback and active learning, challenges included cognitive overload, reduced data immersion, and the need for strong prompting skills. The research emphasizes that empowering educators to design AI tools is crucial for meaningful and pedagogically aligned AI integration in higher education.
In an era where generative AI (Gen-AI) tools are becoming increasingly common in education, a new study sheds light on how educators can actively shape the design and use of these powerful technologies. This research, conducted by Qian Huang and Thijs Willems from the Singapore University of Technology and Design, explores the integration of custom GPT tools into a Master’s-level Qualitative Research Methods course for Urban Planning Policy students.
The study addresses two significant gaps in current AI in education research: the common perception of students as passive AI users, and the limited application of AI in qualitative research methods education. Unlike quantitative disciplines where AI is more frequently used for data analysis, qualitative research, with its emphasis on interpretation, reflexivity, and contextual sensitivity, has seen less AI integration.
Drawing on the Technological Pedagogical Content Knowledge (TPACK) framework and an action research methodology, the instructors designed four custom GPTs. These tools were crafted to support various tasks crucial to qualitative research, such as formulating research questions, practicing interview techniques, analyzing field notes, and applying design thinking principles. The goal was to align Gen-AI with pedagogical intent, ensuring it supports disciplinary learning effectively.
The four custom GPTs developed were:
QualiQuest Buddy GPT
This tool helps students refine research questions, explore concepts like epistemology, ontology, and positionality, and generate topic ideas.
Research Interview Simulator GPT
This GPT simulates interviews, offering adaptive responses, presenting ethical dilemmas, and providing techniques for interview practice.
Observation Station GPT
Designed to guide students in analyzing field notes, helping them distinguish between observation, interpretation, and reflection.
Also Read:
- AI Literacy Shapes Student Experience with AI-Powered Assessments, Study Finds
- AI in Academia: How ChatGPT Can Help Universities Create Better Quizzes
DT X Urban Studies GPT
This tool assists students in applying the Design Thinking Double Diamond Framework to urban challenges.
The researchers collected data through student reflections, AI chat logs, and final assignments. The thematic analysis of this data revealed several key benefits. Students reported enhanced reflexivity, improved interview techniques, and better structured analytical thinking. The AI tools provided quick, personalized feedback, which students found highly valuable for iterative learning, allowing them to refine their approaches without delay. The Research Interview Simulator, in particular, was praised for helping students practice and improve their questioning strategies before real-world interviews.
However, the study also highlighted challenges. Some students experienced cognitive overload due to the sheer volume of AI-generated feedback. Concerns were also raised about a potential reduction in deep immersion with qualitative data, suggesting that while AI is efficient for synthesis, it shouldn’t replace manual, reflective analysis. The quality of AI responses was found to be highly dependent on the student’s prompting skills, emphasizing the need for AI literacy training. Technical limitations, such as usage limits with free GPT versions, also posed initial hurdles, which were later addressed by providing paid access.
A notable finding was the emotional disconnect some students felt, describing AI responses as too mechanical or formulaic, lacking the nuance of human interaction.
The study offers three crucial insights for the future of AI in education. Firstly, AI can be a powerful scaffold for active learning, especially when paired with human facilitation. It augments, rather than automates, teaching. Secondly, custom GPTs can serve as cognitive partners, fostering an interactive learning environment where AI responds to evolving ideas. This human-AI collaboration requires critical human oversight to evaluate and interpret AI outputs. Lastly, educator-led design is critical for pedagogically meaningful AI integration. By empowering instructors to design and customize AI tools, these technologies can be deeply aligned with disciplinary knowledge and ethical values.
This research underscores the importance of shifting the focus from students as passive AI users to educators as active designers. It advocates for institutions to invest in training and resources that enable faculty to confidently build and deploy AI tools, ensuring that AI becomes a dynamic partner in fostering critical, creative, and collaborative learning environments. For more details, you can read the full research paper here.
