Human-AI Partnerships to Support Collaborative Learning

This event is being organized by the NCME Artificial Intelligence in Measurement and Education (AIME) SIGIMIE.

Assessing and supporting learning in collaborative tasks can be challenging because the activities are inherently dynamic, multimodal and multiparty. This talk will discuss approaches and challenges in assessing multimodal collaborative data, such as speech, activity logs, and gestures occurring in noisy classrooms. The talk will further explore how to measure student-student and student-tutor interactions and develop tools leveraging human-AI synergies to support learning. Examples will be provided from two projects. The first develops AI partners to support small groups in real-world classrooms, enabling student knowledge construction, sharing, and uptake, and working through uncertainty during STEM learning. The second supports human math tutors in facilitating rich discussions, fostering relationships among students and tutors, and providing feedback and guidance. Finally, we’ll discuss ethical and practical implications of integrating AI in classroom assessments, emphasizing the need for transparent, equitable practices that incorporate student agency.

Presenter:

  • Peter Foltz, University of Colorado, Boulder
When:  Nov 13, 2024 from 04:00 PM to 05:00 PM (ET)

Location

Online Instructions:
Url: http://us02web.zoom.us/j/81728443502?pwd=aGtPb0x0NGhTYVFPYzl2L2xaZEdGUT09
Login: Meeting ID: 817 2844 3502 Passcode: 649132