Someone to talk to: Using automated characters to support simulated learning activities

Falconer, Liz; Frutos-Perez, Manuel

Abstract

The University of the West of England (UWE) has a large number of students who will pursue subsequent careers in a wide range of professional fields such as engineering, law, business, nursing, teaching, psychology, criminology and design. An important part of that education is the ability to relate theory to practice (Barrett 2011), and developments in technology over the past years have now created opportunities to enable students to experience simulations of events and situations that are difficult, unethical or impossible to organise in the real world, before they put their skills into practice in the real world (Newland 2008). Virtual worlds are proving to be popular and effective environments at UWE for simulations of a range of experiences, such as accident investigations, risk assessments, business ethics cases, psychotherapy practice and sociological experiments. However, as the number of students undertaking these simulations increases, so the call on tutor time will significantly increase. These simulations require to be scalable, to enable their potential for study by large cohorts of students. This year we have experimented with automated non-player characters, also known as “bots,” to enable students to undertake some dialogue during the simulated scenarios without the need for a number of tutors to be available to take particular roles. The bots are currently unsophisticated keyword recognition systems, but even these have proven to have some value in two of the simulations; the accident investigation and the risk assessment, where students were able to gather information from characters they could “talk” to, making more realistic the experience of exploring the environment where the simulations were taking place. This paper discusses the results of student feedback, evaluations of these simulations and prototype development for the next generation bots that we want to implement in future learning simulations based on the findings of the evaluations.

• Journal Article Type: Article
• Publication Date: Nov 10, 2011
• Deposit Date: Apr 3, 2012
• Publicly Available Date: Nov 15, 2016
• Journal: 10th European Conference on E-learning
• Peer Reviewed: Peer Reviewed
• Pages: 170
• Keywords: learning simulations, non-player characters, professional practice, virtual worlds
• Public URL: https://uwe-repository.worktribe.com/output/957680
• Publisher URL: http://academic-conferences.org/pdfs/ECEL_2011-Booklet.pdf
• Related Public URLs: http://academic-conferences.org/ecel/ecel2011/ecel11-home.htm
• Contract Date: Nov 15, 2016