Publikation

Titel:

Human-AI Decision Support Interaction: Essays on Trust and Reliance

AutorInnen:

Strunk, Jobin

Kategorie:

Dissertationen

erschienen in:

Dissertation, FernUniversität in Hagen, Hagen, 2026. {Link}

Abstract:

This thesis investigates the multifaceted phenomenon of trust in artificial intelligence (AI)-based decision support systems (AI-DSSs) and examines how trust shapes human interaction, reliance, and delegation in increasingly agentic information systems (IS). The overarching research objectives of this thesis are threefold. First, it seeks to clarify and conceptualize the core constructs underlying trust in human-AI interaction, including cognitive and emotional trust, delegation, agency, and contextual risk. This conceptual grounding is essential for understanding how AI-DSS differ from traditional IS and why trust plays a pivotal role in their adoption and use. Second, the thesis examines how AI-DSS characteristics, communication strategies, and contextual conditions influence trust formation and calibration, thereby explaining when and why users choose to rely on AI-DSS advice. Third, it aims to derive actionable design principles that enhance perceived trustworthiness in AI-DSS, bridging the gap between theoretical insight and practical development.

The research findings are structured into five cumulative essays, each addressing different aspects of trust in AI-DSS and collectively answering the thesis’s overarching research questions. This introductory essay presents the motivation for studying trust in the context of increasingly agentic IS, outlines the conceptual and theoretical foundations, and formulates the research questions that guide the thesis. Moreover, it describes the research process and explains how the five essays ranging from conceptual synthesis and experimental studies to a design science project jointly contribute to a comprehensive understanding of trust and its role in shaping effective and responsible human-AI interaction.