A Two-Essay Examination of Human-AI Collaboration in Creative Ideation

Abstract

Generative artificial intelligence (GenAI) is increasingly embedded in creative work, yetresearch has not adequately explained how human-AI collaboration shapes both the human experience of creative work and the process through which users influence AI-generated outcomes. This dissertation examines human-AI collaborative ideation through two essays that share a common premise: the value of GenAI depends not only on technology’s capability but also on whether human agency remains consequential during collaboration. Essay One develops a dual-path model explaining how GenAI can both activate and detract from earned dignity in creative ideation. Using an online experiment with a self- developed AI-supported ideation platform, Idea Hub, the essay shows that GenAI’s breadth and depth of exploration increase AI-enabled serendipity, which in turn improves perceived ideation performance and reinforces earned dignity. Moreover, AI-induced fixation undermines dignity primarily by increasing perceived dispensability. Essay Two focuses on the interaction process by introducing the concept of contextual steering behavior (CSB), defined as users' deliberate injection of contextual, constraint-based, and future-oriented knowledge into the ideation process. CSB is conceptualized through three subdimensions: experiential grounding (drawing on prior experience), practical bounding (introducing contextual constraints), and projective recasting (redirecting outputs toward anticipated future needs). Essay Two also develops a research model of human-AI collaborative ideation that explains how collaboration strategies shape creative outcomes through user contextual steering behavior. The research model is largely supported by an online experiment.