People from all over the world must live and work together in today’s society. Such integration is not always a smooth process, and interactions with people from other cultures may lead to misunderstandings or even outright conflicts.
In the last few years, researchers and practitioners have been working on creating digital tools that can be used to mediate these misunderstandings and conflicts. These tools typically involve interactions with so-called intelligent agents, i.e. virtual characters that are able to take decisions autonomously, that behave as if they are from another culture. The aim of these interactions is to make potential trainees experience how misunderstandings can shape interactions with and perceptions of people from other cultures.
In this work, we take the first steps in the design of a digital culture-general training tool to help young adults deal with misunderstandings or conflicts due to differences in culture, through interactions with intelligent agents. We have posed the following design research questions and found the following answers:
Which concepts are required to describe the design of a digital culture-general training tool involving agents that show culturally varying behaviour?
The answer to this question can be found in the glossary, which presents the key concepts that have been used in this work to create agents that show culturally varying behaviour and to create scenarios that incorporate these agents to increase the intercultural competence of trainees.
Can we use theories of culture to create scripted scenarios in which virtual characters behave appropriately for a given culture?
To answer this question we designed scripted scenarios in which virtual characters show culturally varying behaviour based on a theory of culture. To ensure that the behaviours of these virtual characters were representative of real-life cultural differences, we conducted an evaluation with people from a wide range of cultures. The results show that the dimensions of culture can be used to generate culturally varying behaviour in agents, but that extensive (pre)testing is required to ensure that the underlying intention of the characters’ behaviour aligns with the users’ interpretation of that behaviour.
Can we identify requirements for sociocultural agents that can help them to make sense of their social world?
To answer this question we focused on describing important concepts of social interactions based on theories from sociology and psychology. These concepts are incorporated into a conceptual model for socio-cultural agents that can be used to describe their social world. The model differentiates between three levels of analysis: the interaction, the group, and the society. These levels range from being more specific, and thus more visible, to more abstract, and thus less visible, and help us to understand how each level affects interpretation and behaviour.
Can we create intelligent agents that can vary their behaviour depending on the culture to be simulated?
To answer this question we described the creation of intelligent agents that show culturally varying behaviour. We use an existing model to create believable social interactions, in which agents attribute, claim, and confer social importance in their interactions with other agents and users. Social importance is a way to measure the importance of a certain individual in the eyes of others. The strength of attribution, claims, and conferrals was varied using cultural modifiers. The generated behaviour of the agents was then evaluated to ensure that the intelligent agents showed behaviour representative of a given culture. The results suggest that it is possible to create intelligent agents that can act out appropriate culturally varying behaviour for a given culture.
Can we create critical incidents, involving intelligent agents that show appropriate behaviour for given cultures, through which potential trainees become more sensitive to and knowledgeable about differences across cultures?
To answer this question we focused on applying different methods of intercultural training in the design of a digital culture-general training tool. These methods were incorporated into critical incidents, in which users can interact with intelligent agents. To ensure that the critical incidents led to an attribution of perceived differences in behaviour to specific differences in culture and to (potential) trainees becoming less judgemental of inappropriate behaviours by people from different cultures, the tool was evaluated by two groups of students. The results suggest that it is possible to create agent-based critical incidents to make potential trainees more knowledgeable about differences across cultures.
The findings to our design research questions represent a set of important contributions to the field.
First, we have identified and structured important concepts to better understand the design and implementation of socio-cultural agents and the design of critical incidents that involve these agents for intercultural training. Second, we have described and used models that help to define the simulated world of the agents and help them to navigate through that world. Third, we have attempted to systematize the process of creating scenarios involving agents that show culturally varying behaviour through a set of guidelines that need to be met to ensure that the behaviour of socio-cultural characters is properly evaluated. Fourth, besides conceptual elements, we have also created practical implementations that can freely be used and modified by others.
In our work, we have only taken the first steps in designing a digital culture-general training tool. Additional work on the generalization and validation of the critical incidents and the behaviours of the agents is still required; however, we believe that our results show our approach to be viable. We believe that future work will have to focus on three fields: understanding how trainees can be emotionally engaged in the scenarios, systematizing the process of using model-driven approaches to generate socio-cultural behaviour, and using the design outputs in different contexts and with different people from different cultures.
|Qualification||Doctor of Philosophy|
|Award date||10 Dec 2014|
|Place of Publication||Wageningen|
|Publication status||Published - 2014|
- information technology
- intercultural communication