Abstract

Generative artificial intelligence (GenAI) agents have rapidly evolved from assistive tools into systems embedded in employees’ work and decision-making processes (Benbya et al., 2024). Because GenAI can instantly generate structured and realistic responses, employees increasingly rely on it for information search, decision support, and problem-solving. However, AI-generated content may also contain factual errors, biases, reasoning inconsistencies, or fabricated information, commonly referred to as AI hallucinations (Wei et al., 2024). Since such responses are often delivered with high fluency and confidence, employees may accept them without recognizing potential inaccuracies, potentially leading to poor decisions and negative organizational outcomes. Although employees sometimes verify AI-generated information through external sources or additional AI tools, existing research mainly focuses on adoption intention, continuance usage, and trust perceptions, with limited attention to verification behavior after receiving AI-generated outputs. To address this gap, this study integrates the Elaboration Likelihood Model (ELM) and the Motivation–Opportunity–Ability (MOA) framework to explain how employees evaluate AI-generated content and translate skepticism into verification behavior. In this study, skepticism toward AI-generated content represents motivation for verification, while verification ability and time pressure determine whether skepticism leads to actual behavior. This study employs a 2×2×2 between-subjects experimental design using a literature review task that simulates AI hallucinations, such as fabricated citations and misleading conclusions. AI content quality and interaction cues function as central and peripheral cues within the ELM framework, while verification behavior is captured through system logs under varying levels of verification ability and time pressure. By integrating ELM and MOA, this study explains how employees move from cognitive evaluation of AI-generated content to verification behavior.

Download

Share

COinS