The effects of race and authenticity on face identification: a study on the majority race in a multiracial society

• Main Author: Hang, Yi Zhen
• Other Authors: Charles Or
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Social sciences::Psychology
• Online Access: https://hdl.handle.net/10356/172822
• Institution: Nanyang Technological University

In multiracial Singapore, where the Chinese constitute the majority, investigating own-race bias becomes pertinent. Own-race bias involves a better recognition of faces from one's own race. Additionally, given the growing realism of artificial intelligence (AI)-generated faces, it is pertinent to investigate individuals' ability to distinguish between AI-generated and real faces. Hence, this study examined the effects of authenticity and race on face identification and face identification for AI-generated and real faces through a series of tasks. Specifically, participants engaged in face memory tasks with six conditions, alongside race and authenticity identification tasks. In each face memory task condition, 15 faces were shown twice, subsequently identifying them among 30 faces (15 seen, 15 distractors). After all six face memory tasks, participants engaged in race and authenticity identification tasks, where all 180 faces from the six conditions were presented. Participants were tasked with identifying each face's race and authenticity. Two versions of the task were administered, one presenting 180 faces with their original brightness and the other displaying the same faces with adjusted brightness. The results revealed that own-race bias was only observed specifically between AI-generated Chinese and AI-generated Malay faces, while it was not evident in the case of real Chinese, Indian, and Malay faces or between AI-generated Chinese and AI-generated Indian faces. Moreover, the study found that participants struggled to differentiate between AI-generated faces and real faces. These findings contribute valuable insights into the intricate dynamics of face recognition, particularly concerning the interaction between AI-generated faces and real faces.