What question did this study set out to answer?

The aim is to explore how trust in AI manifests in human-AI relationships, particularly in healthcare and labor contexts.

June 3, 2026Open Access

Rethinking trust in AI: empirical insights from labor and healthcare domains

Key Points

The aim is to explore how trust in AI manifests in human-AI relationships, particularly in healthcare and labor contexts.
Conducted two empirical studies focusing on public perceptions and AI communication in healthcare and labor.
Study 1 gathered data on AI capability and social desirability across 100 occupations from a U.S. sample.
Study 2 examined influences of diagnostic uncertainty types communicated by medical AI chatbots on perceived credibility.
Results indicate trust in AI is influenced by perceptions of competence and communication style.
Findings reveal that functional and relational aspects of trust are context-dependent in both domains.
The studies highlight significant implications for AI system design and enhancing user engagement.

Abstract

Abstract This manuscript explores the evolving concept of trust in human-AI relationships through two empirical studies conducted in complementary domains. As artificial intelligence (AI) becomes increasingly integrated into high-stakes decision-making domains such as healthcare and employment, we examine how trust manifests along two key dimensions: functional (related to capability and performance) and relational (related to social desirability and sincerity). Study 1, based on previously published analyses (Oh & Jung, 2023), examines public perceptions of AI's technological capability and social desirability in replacing human workers across 100 occupations, drawing on responses from a U.S.-based sample. Study 2 explores how different types of diagnostic uncertainty, epistemic and aleatory, communicated by medical AI chatbots affect perceived credibility and sincerity. Together, the findings underscore that trust in AI is deeply context-dependent, influenced by perceived competence, communication style, and user expectations. The manuscript contributes to understanding trust as involving both functional and relational dimensions in human–AI interaction, with implications for AI system design and user engagement.

Ask AI

Helpful

Bookmark

View Full Paper