Work-in-Progress—Virtual Agents in Teaching: A Study of Human Aspects

Abstract

Students require human intelligence and social interaction in the form of academic assistance at different times of their study period. Their desire to get and find academic assistance varies and is dependent on many factors such as attendance mode, personal situation, semester timetables, and assessment due dates. Providing students with access to this expertise when it is needed and to large numbers of students is problematic. Virtual Agents (VAs) seek to provide a technology-enabled social element to encourage and provide timely support to aid students' learning. We have implemented 4 unit-specific VIRtual Teaching Assistants (VIRTAs) across 2 universities to provide support to answer student's questions about various aspects of the unit. In this paper, we present the usage patterns of students to show how many questions were asked by students and at what point of time in the semester the questions were asked addressing the desire to find assistance when required from VIRTA.

A proposal to measure the impact of automated response systems on meeting student learning outcomes

“…As noted in the introduction, AI has been in use for online learning environments to act as a teaching aide or as an academic support, such as 'Jill' Watson, a knowledge based artificial intelligence (KBAI) virtual teaching assistant (Maderer, 2016) and VIRtual Teaching Assistant (VIRTA) (Jha et al, 2020), but limited mainly to functionality similar to that of a marketing-style chatbot (such as chatbots that recommend products to customers in commerce shopping sites). This more functional approach of Automated Response Systems (ARS) acts more as a directory and may not be useful in supplementing teachers in a more dynamic settings such as online learning environments.…”

“…ARS are pre-designated responses or replies that are generated by a software program for incoming queries and/or messages. Commercially available ARS systems end to have a combination of features highlighted, such as personalised responses, instant feedback, high accessibility, 24/7 support and the ability to provide basic feedback (Jha et al, 2020;Microsoft, 2020;Pinkwart, 2016;Squirrel AI, 2019). In place of supporting purchasing and marketing, ARS can be implemented (especially in online learning environments) to support how students learn.…”