Options
Human-Machine Interaction at the Intersection of Artificial Intelligence and Cognitive Psychology: Exploring Consumers` Aversion towards AI
Type
doctoral thesis
Date Issued
2022-02-21
Author(s)
Abstract
Despite the performance superiority of artificial intelligence (AI) in a wide range of application areas, many users do not accept AI-based solutions. The phenomenon of AI aversion has been widely addressed across more than a decade of research, yet it remains unclear why consumers react aversely to AI solutions. Due to the identified tension, this dissertation addresses current and future human-machine interaction at the intersection of AI and cognitive psychology in three independent research projects: The first project compares consumer perceptions of AI-created art with human crafted art. The results demonstrate that consumers assign a lower monetary and aesthetic value to technology-based artworks once the contribution of AI becomes apparent. The second project applies the theory of emotion-based decision-making as a starting point to investigate the decision mechanism for or against AI-based task completion. The willingness of humans to delegate a task to AI decreases with rising level of a positive and arousing emotional state, which is evoked by the task. The third project illustrates the fundamentally different perception and understanding of consumers concerning AI itself, as well as AI-related terminologies. The way subjects evaluate AI is classified into tangible and intangible representations of AI (e.g., consumers assess robot as a tangible and algorithm as an abstract entity). Overall, the three projects depict fundamentals for acceptance when introducing AI-based solutions. In addition, the findings support practitioners in the acceptance assessment of AI-based solutions in a business context. From a methodological standpoint, the application of interdisciplinary research techniques, such as image style networks, contributes to the usage of deep learning models as an unexplored method for developing stimuli in consumer research.
Abstract (De)
Despite the performance superiority of artificial intelligence (AI) in a wide range of application areas, many users do not accept AI-based solutions. The phenomenon of AI aversion has been widely addressed across more than a decade of research, yet it remains unclear why consumers react aversely to AI solutions. Due to the identified tension, this dissertation addresses current and future human-machine interaction at the intersection of AI and cognitive psychology in three independent research projects: The first project compares consumer perceptions of AI-created art with human crafted art. The results demonstrate that consumers assign a lower monetary and aesthetic value to technology-based artworks once the contribution of AI becomes apparent. The second project applies the theory of emotion-based decision-making as a starting point to investigate the decision mechanism for or against AI-based task completion. The willingness of humans to delegate a task to AI decreases with rising level of a positive and arousing emotional state, which is evoked by the task. The third project illustrates the fundamentally different perception and understanding of consumers concerning AI itself, as well as AI-related terminologies. The way subjects evaluate AI is classified into tangible and intangible representations of AI (e.g., consumers assess robot as a tangible and algorithm as an abstract entity). Overall, the three projects depict fundamentals for acceptance when introducing AI-based solutions. In addition, the findings support practitioners in the acceptance assessment of AI-based solutions in a business context. From a methodological standpoint, the application of interdisciplinary research techniques, such as image style networks, contributes to the usage of deep learning models as an unexplored method for developing stimuli in consumer research.
Language
English
Keywords
Künstliche Intelligenz
Kognitive Psychologie
Automation
EDIS-5165
Cognitive Psychology
Artificial Intelligence
HSG Classification
not classified
HSG Profile Area
None
Publisher
Universität St. Gallen
Publisher place
St.Gallen
Subject(s)
Division(s)
Eprints ID
265925
File(s)