Designing for Continuous Interaction with Artificial Intelligence Systems

P Wintersberger, Berkel N Van, N Fereydooni, B Tag, EL Glassman, D Buschek, Ann Blandford, F Michahelles
in CHI EA '22: CHI Conference on Human Factors in Computing Systems, Conference paper (text)


The increasing capabilities of Artificial Intelligence enable the support of users in a continuously growing number of applications. Current systems typically dictate that interaction between user input and AI output unfolds in discrete steps, as is the case with, for example, conversational agents. Novel scenarios require AI systems to adapt and respond to continuous user input, e.g., image-guided surgery and AI-supported text entry. In and across these applications, AI systems need to support more varied and dynamic interactions in which users and AI interact continuously and in parallel. Current methods and guidelines are often inadequate and sometimes even detrimental to user needs when considering continuous usage scenarios. Realizing a continuous interaction between users and AI requires a substantial change in perspective when designing Human-AI systems. In this SIG, we support the exchange of cutting-edge research contributing to a better understanding and improved methods and tools to design continuous Human-AI interaction.