hide
Free keywords:
Evaluation; Intelligent Virtual Agent; Enactment; Training;
Learning; Foreign Language
Abstract:
Virtual (animated software) agents can train humans in
vocabulary learning. This has been successfully tested with adults and
more recently also with children. However, the question of how children
perceive a virtual agent training them had not been investigated. Here
we invited 25 children to evaluate their perception of a virtual and a
human trainer who presented written words in a foreign language on
videos; both the human trainer and the virtual agent additionally
performed a semantically related gesture for each word. Subjects rated
the trainers for features related to gestures and for their “personalities”.
Subjects found human gestures better and gave the human trainer
higher sympathy scores; however, the overall difference between their
perception of virtual and human trainers was not significant.
