An Analysis of Human-AI Interaction

Abstract

The pervasive integration of AI technologies, such as ChatGPT, into various sectors of daily life has necessitated a shift in AI research. While initial studies primarily concentrated on technological advancements, the current focus has transitioned towards a human-centric approach in understanding AI. This special issue aims to expand the body of knowledge on human-AI interaction, addressing the challenges presented by AI's increasingly human-like capabilities and the diminishing distinction between individuals and technology.

The issue introduces a comprehensive framework for human-AI interaction, conceptualizing AI as an agent for human users. Two critical types of compatibility are identified: human-AI fit and task-AI fit. Additionally, the framework considers varying degrees of AI agency, resulting in a human-AI collaboration continuum. This conceptualization leads to a research framework that examines AI characteristics, behaviours, and outcomes, moderated by user and task characteristics.