WIP: An AI-Based Virtual Assistant for Supporting a Large Engineering Course

Abstract

The work-in-progress innovative practice paper showcases the development of an AI-based virtual Teaching Assistant (TA) in a large freshmen engineering course with 800+ students per semester. In this environment, it is challenging for the instructional team to resolve all the administrative issues students may encounter, provide individualized support, and answer questions on time. If not handled efficiently, these environments can easily result in frustration both for students and the instructional team. Repetitive time-consuming tasks often manifest in the form of emails and help form submissions, which typically have a turnaround time of 1–2 days depending on the volume and represent a significant workload for the instructional team. To address administrative questions related to the course, including assignments, software issues, and grading, we trained and deployed a custom virtual assistant based on ChatGPT 4.0. Our system provides quick and concise support to students who need guidance, clarification, and further explanation of the technical topics covered in the course. The system significantly releases TAs' time, allowing them to focus on direct instruction and student interaction.

Ähnliche Arbeiten