Risk perception of biased AI-enabled technologies in healthcare: investigating the role of negative cognition, negative emotion, and reactance

• Main Author: Yi, Sue Hyon
• Other Authors: Andrew Prahl
• Format: Thesis-Master by Research
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Social Sciences; Risk communication; Health communication; AI-enabled technologies; Reactance; Message framing
• Online Access: https://hdl.handle.net/10356/174204
• Institution: Nanyang Technological University

The worldwide implementation of AI-enabled technologies has significantly altered the current healthcare environment. Current raised issues about biased AI technologies have increased risk perception. In order to rectify the negative perception towards biased AI technologies in healthcare, the effect of framing (i.e., gain and loss frame) need to be investigated. There has been a gap in the risk communication literature dealing with newly emerging technologies that are embraced with uncertainties about the role of negative cognition and negative emotion on the two dimensions of risk perception (i.e., perceived susceptibility and perceived severity). This study conducted an online experiment to assess the effects of framing, the role of negative cognition and negative emotion in interacting with message framing, and explore the mediating mechanisms of risk perception. The results revealed that the gain frame was effective in reducing perceived severity and perceived susceptibility of biased AI-enabled technologies as opposed to the loss frame among the US public. Also, negative cognition and negative emotion interact with message frames in predicting perceived severity. Moreover, negative cognition and negative emotion significantly mediate the relationship between the threat to freedom and risk perception. This study contributes to diversifying the effects of gain and loss framing, and enlightening the process of risk perception by exploring the role of negative cognition and negative emotion.