Multicriteria Decision Analysis and Conversational Agents for Children with Autism

Abstract

Conversational agents has emerged as a new means of communication and social skills training for children with autism spectrum disorders (ASD), encouraging academia, industry, and therapeutic centres to investigate it further. This paper aims to develop a methodological framework based on Multicriteria Decision Analysis (MCDA) to identify "the best", i.e. the most effective, conversational agent for this target group. To our knowledge, it is the first time the MCDA is applied to this specific domain. Our contribution is twofold: i) our method is an extension of traditional MCDA and we exemplify how to apply it to decision making process related to CA for person with autism: a methodological result that would be adopted for a broader range of technologies for person with impairments similar to ASD; ii) our results, based on the above mentioned method, suggest that Embodied Conversational Agent is most appropriate conversational technology to interact with children with ASD.