Building a Teacher-AI Collaborative System for Personalized Instruction and Assessment of Comprehension Skills
PERIOD:
TO
Dr. Ying Xu (PI) and colleagues will leverage generative artificial intelligence (AI) and large language models (LLMs) to empower teachers to co-create STEM-focused reading resources with AI. These resources feature an adaptive chatbot as an intelligent reading partner that asks students questions, listens to and interprets the student responses, and provides tailored feedback to students during reading.
The teacher-AI collaboration mechanism allows teachers to oversee the question generation process and customize the chatbot question sequence, aligning it with their instructional objectives and the specific needs of students. This project will be carried out in four stages. First, the team will develop innovative AI models to automatically generate question-answer pairs based on reading materials teachers select, and tailor the AI models to meet the unique requirements of an educational context. Second, through a contextual inquiry and participatory design process, the team will develop a user-friendly teacher-AI collaborative system that allows teachers to easily verify and modify the question-answer pairs generated by AI and to subsequently generate chatbots that can engage students in dialogue. Teachers’ modifications will feed back to the system so that the AI models can gradually learn and adapt to each individual teacher’s preferences. Third, the team will develop the chatbot’s capability for adaptive interaction so that it can carry out dialogue and provide scaffolding based on both the accuracy and sentiment of students’ responses. Fourth, the team will examine the usability and effectiveness of the teacher-AI collaborative system and the resulting chatbot in supporting personalized instruction and formative assessment in an under power randomized controlled trial. Observations and interviews with teachers and students will shed light on the usability of the teacher-AI co-created interactive reading materials. Students’ post-reading comprehension will be assessed to provide evidence on the system’s educational impact. This project is among the first to explicitly focus on scaling up the development of AI-based interactive reading materials to ensure that these resources reach and benefit the broadest possible range of learners. It also promises to offer a generalizable case of non-technical domain experts being deeply engaged in the generation of AI-based learning resources.
Other co-PIs include Mark Warschauer, PhD, University of California-Irvine, Young-Suk Grace Kim, EdD, University of California-Irvine, Shiyu Chang, PhD, University of California-Santa Barbara, and Dakuo Wang, PhD, Northeastern University.
