Exploring Adaptation of Conversational Systems to Different Age Groups
Ideally, conversational agents should be able to adapt to human users. However, state-of-the-art AI systems currently lack sophisticated adaptation capabilities. Our long-term goal is to develop artificial agents that can adapt to individuals/user groups at any level (age, expertise, language style, etc.) and that are perceived as trustable by users. The current project focuses on users of different age groups, and asks: (1) Can some degree of adaptation be achieved by training a state-of-the-art system with data targeted to specific age groups? (2) Does this lead to differences in the perceived degree of anthropomorphism, social presence, trust, appreciation and comprehensibility of the message? We will implement an open-source demo showcasing the results of the project, and findings will also be disseminated through blog posts and a hands-on workshop.
Dr. Margot van der Goot is Assistant Professor at the Amsterdam School of Communication Research (ASCoR), Department of Communication Science. Her research focuses on users’ perceptions of interactions with conversational agents.
Dr. Raquel Fernández is Associate Professor at the Institute for Logic, Language and Computation (ILLC), Faculty of Science. She is an expert on modelling dialogue by humans and machines using methods from computational linguistics.
Dr. Sandro Pezzelle is Postdoctoral researcher at the Institute for Logic, Language and Computation (ILLC). His research is on multimodal, cognitively-inspired AI models for language understanding and generation.