Designing human-agent interaction for AI literacy of novice adult learners
Patel, Devangini (2023)
Patel, Devangini
2023
Master's Programme in Computing Sciences
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2023-05-23
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202304234145
https://urn.fi/URN:NBN:fi:tuni-202304234145
Tiivistelmä
In the coming few years, at least 50% of the workers would need to upskill for AI-based jobs such as data analysts, process automation specialists, and digital transformation specialists. Currently, existing AI courses are not suitable for novice adult learners learning online as these courses require technical knowledge, do not involve practical activities, or require teaching assistance. Previous user studies with practical AI literacy systems found students needed assistance in selecting input data, iteratively building AI, and measuring AI’s performance. A pedagogical agent (computer program to assist in multimedia teaching) could provide this assistance to novice adult learners. A limited number of AI literacy pedagogical agents for novice adult learners in an online context have been designed and evaluated. The goal of this thesis is to identify: (1) the user needs of novice adult learners, (2) user experience (UX) goals for an AI literacy system to be used by novice adult learners in an online self-study context, and (3) human-agent interaction that satisfies user needs and achieve UX goals. This thesis followed an Experience-Driven Design (EDD) approach; the user needs of novice adult learners in AI literacy in an online context were identified, and then human-agent interaction to aid novice adult learners in AI literacy in an online context was designed and evaluated. A user needs study with 8 participants helped to identify the UX goals of relatedness, engagement, and competence; and it revealed that adult learners need a social agent to provide emotional- and task-based feedback to guide them in the correct direction. Two human-agent interactions were iteratively designed based on feedback types: (1) task feedback, and (2) emotional and task-based feedback. An evaluation study with these agent feedback types was conducted with 8 participants using a between-subject design. The results showed that a pedagogical agent for AI literacy in an online context should (1) provide immediate emotions to communicate AI confidence in real-time, (2) use sounds to provide data input feedback, and (3) answer task-related questions.