I’m pleased to share that I will present our accepted work at the Singapore International Science Teachers’ Conference (SISTC) 2026, taking place at Science Centre Singapore from 24 to 26 November 2026.
The revised title is “From Misconceptions to Evidence: What Science Teachers Make Visible When Co-Designing Agentic Learning Apps.”
The work is co-authored with Liow Wei Ting, Sumbul Khan, and Ang Lay Kee.
What the study examines
Science teachers do more than deliver correct answers: they elicit learners’ models, diagnose misconceptions, interpret evidence, and preserve professional judgment. Our study examines how teachers translate that epistemic work into specifications for agentic learning applications.
Across four de-identified teacher-created artifacts, we found that teachers consistently connected a science-learning problem to a learner interaction and to evidence that could inform teaching. The proposals positioned AI less as an answer generator than as an elicitor, scaffold, and evidence-return mechanism. At the same time, teacher authority and safeguards were unevenly specified.
We therefore propose a five-question design protocol for accountable educational AI: What is the problem? What will the learner do? What evidence will become visible? What authority remains with the teacher? What safeguard is required?
I look forward to discussing the work with the SISTC community. I will share the specific session date and time when the conference programme is released.
This work acknowledges support from the Design-centric Interdisciplinary Creative Problem Solving (ICPS-D): Developing Critical and Creative Skills in Teachers project.
Leave a Reply