Loading...
Paper Number
2523
Paper Type
Short
Abstract
The emergence of generative artificial intelligence has revolutionized various sectors by providing automated solutions. However, the lingering challenges posed by AI hallucinations cannot be ignored. While existing research explores critical facets of AI-human collaboration, the phenomenon of AI hallucinations remains underexplored. In particular, there are significant gaps in how the characterization of AI-generated information affects users' detections of AI hallucinations. In this conceptual paper, drawing on Heuristic-Systematic models of information processing, we explore how the structure, emotional tone, and mode of presentation of the information affect the detection of AI hallucination and propose a dual-process hallucination detection model. This study will extend the literature on AI-human collaboration and enrich the understanding of cognitive processes underlying the identification of AI hallucinations. Additionally, the findings of this paper will have significant implications for the design of AI interfaces, which can enhance the integration of AI into workplace processes.
Recommended Citation
Tao, Yanda; Yoo, Changseung; and Animesh, Animesh, "Detection of AI Hallucinations: The Impact of Information Characteristics" (2024). ICIS 2024 Proceedings. 5.
https://aisel.aisnet.org/icis2024/aiinbus/aiinbus/5
Detection of AI Hallucinations: The Impact of Information Characteristics
The emergence of generative artificial intelligence has revolutionized various sectors by providing automated solutions. However, the lingering challenges posed by AI hallucinations cannot be ignored. While existing research explores critical facets of AI-human collaboration, the phenomenon of AI hallucinations remains underexplored. In particular, there are significant gaps in how the characterization of AI-generated information affects users' detections of AI hallucinations. In this conceptual paper, drawing on Heuristic-Systematic models of information processing, we explore how the structure, emotional tone, and mode of presentation of the information affect the detection of AI hallucination and propose a dual-process hallucination detection model. This study will extend the literature on AI-human collaboration and enrich the understanding of cognitive processes underlying the identification of AI hallucinations. Additionally, the findings of this paper will have significant implications for the design of AI interfaces, which can enhance the integration of AI into workplace processes.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
10-AI