Jiayi Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jiayi Zhu has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Zituo Wang
University of Southern California, Los Angeles, CA, USA
4 shared publications - Donggyu Kim
Annenberg School for Communication and Journalism, University of Southern California
2 shared publications Jianyu Wu
Wuhan University, Wuhan, China
1 shared publication- Vivian Jiaying Zhai
University of Southern California, Los Angeles, USA
1 shared publication Wenqing Zuo
Nanjing University, China
1 shared publicationYufei Chen
Huazhong University of Science and Technology, China
1 shared publication
Latest Publications
- Open Access
- 8 April 2026
- Published by SAGE Publications in Science Communication
- Vol. 48 (3) , 305-337
Generative AI So White: Racial Biases in AI Imagery Across the United States and China
This study investigates racial biases in AI-generated occupational images across models developed in the United States and China. Situated at the intersection of human–AI communication and postcolonial theory, we conceptualize generative AI as an active participant in science communication that shapes visual knowledge and racial representation within a global postcolonial order. Constructing a dataset of 9,600 images generated by four models (GPT-4o, Llama 3, Wanx2.0, and Wenxin 3.5), we examine three levels of racial biases—representational bias, positional bias, and racialized meaning bias—using a mixed-methods approach. Findings show that White individuals are overrepresented, granted spatial dominance, and encoded through aesthetic and symbolic conventions that racialize non-White bodies. We reveal how generative AI reproduces global racial hierarchies under algorithmic neutrality, advancing a cross-national auditing framework and contributing to decolonial science communication by foregrounding a human-centered AI perspective.
- 8 April 2026
- Published by SAGE Publications in Science Communication
- Vol. 48 (3) , 305-337
Open AccessGenerative AI So White: Racial Biases in AI Imagery Across the United States and China
This study investigates racial biases in AI-generated occupational images across models developed in the United States and China. Situated at the intersection of human–AI communication and postcolonial theory, we conceptualize generative AI as an active participant in science communication that shapes visual knowledge and racial representation within a global postcolonial order. Constructing a dataset of 9,600 images generated by four models (GPT-4o, Llama 3, Wanx2.0, and Wenxin 3.5), we examine three levels of racial biases—representational bias, positional bias, and racialized meaning bias—using a mixed-methods approach. Findings show that White individuals are overrepresented, granted spatial dominance, and encoded through aesthetic and symbolic conventions that racialize non-White bodies. We reveal how generative AI reproduces global racial hierarchies under algorithmic neutrality, advancing a cross-national auditing framework and contributing to decolonial science communication by foregrounding a human-centered AI perspective.
- 12 February 2026
- Published by SAGE Publications in New Media & Society
Justice for @birdy: Queer counterpublics on Chinese TikTok (Douyin)
This study investigates the queer counterpublics on Douyin that call for social justice on LGBTQ rights. By revealing Douyin’s affective affordances to produce queer counterpublics through censorship, we propose a framework that explains how queer counterpublics navigate user–state tensions and balance affect with strategy in a dynamic discursive space. Based on our large-scale Douyin video dataset, we analyze 102,031 videos from 90,901 accounts by combining computational and qualitative methods. By employing computer vision and content analysis, we uncover the multimodal features and strategies calling for justice in those videos. We also show how non-influencers, influencers, and official accounts adopt distinct strategies and how these strategies are associated with user engagement. Further multimodal discourse analysis examines three key counter discourses, including maximum visibility, limited anger, and infinite hope. This study contributes to the understanding of queer counterpublics in China and the potential of short video platforms like Douyin to foster social change.
- 12 February 2026
- Published by SAGE Publications in New Media & Society
Justice for @birdy: Queer counterpublics on Chinese TikTok (Douyin)
This study investigates the queer counterpublics on Douyin that call for social justice on LGBTQ rights. By revealing Douyin’s affective affordances to produce queer counterpublics through censorship, we propose a framework that explains how queer counterpublics navigate user–state tensions and balance affect with strategy in a dynamic discursive space. Based on our large-scale Douyin video dataset, we analyze 102,031 videos from 90,901 accounts by combining computational and qualitative methods. By employing computer vision and content analysis, we uncover the multimodal features and strategies calling for justice in those videos. We also show how non-influencers, influencers, and official accounts adopt distinct strategies and how these strategies are associated with user engagement. Further multimodal discourse analysis examines three key counter discourses, including maximum visibility, limited anger, and infinite hope. This study contributes to the understanding of queer counterpublics in China and the potential of short video platforms like Douyin to foster social change.
- 19 January 2026
- Published by SAGE Publications in Communication & Sport
- Vol. 14 (5) , 1773-1800
Hybrid Identity in Motion: Eileen Gu’s Cross-Platform Visual Self-Presentation on Weibo and Instagram
This study investigates the cross-cultural visual self-presentation of skier Eileen Gu, a U.S.-born athlete who migrated competitively to represent China at the Beijing 2022 Winter Olympics. While previous scholarship has examined how sports media construct nationalistic and gendered narratives around athletes, little attention has been paid to how athletes themselves negotiate these identities through visual self-presentation on social media. Drawing on self-presentation theory, we analyzed 893 images from Gu’s official Weibo and Instagram accounts. Using a combination of human and computational content analysis, we found that Gu’s visual self-presentation highlighted personal life and used more sexual appeals on Instagram, but centered on athletics and national symbols on Weibo. Athletics-themed posts received the highest engagement, while nationalistic symbols boosted likes and comments on Instagram but reduced shares on Weibo. Visual characteristics and facial expressions predicted engagement only on Instagram. By integrating computational methods with self-presentation theory, this study contributes to sports migration scholarship by showing how athletes negotiate hybrid identity and mobilize cross-cultural support in digital sport communication.
- 19 January 2026
- Published by SAGE Publications in Communication & Sport
- Vol. 14 (5) , 1773-1800
Hybrid Identity in Motion: Eileen Gu’s Cross-Platform Visual Self-Presentation on Weibo and Instagram
This study investigates the cross-cultural visual self-presentation of skier Eileen Gu, a U.S.-born athlete who migrated competitively to represent China at the Beijing 2022 Winter Olympics. While previous scholarship has examined how sports media construct nationalistic and gendered narratives around athletes, little attention has been paid to how athletes themselves negotiate these identities through visual self-presentation on social media. Drawing on self-presentation theory, we analyzed 893 images from Gu’s official Weibo and Instagram accounts. Using a combination of human and computational content analysis, we found that Gu’s visual self-presentation highlighted personal life and used more sexual appeals on Instagram, but centered on athletics and national symbols on Weibo. Athletics-themed posts received the highest engagement, while nationalistic symbols boosted likes and comments on Instagram but reduced shares on Weibo. Visual characteristics and facial expressions predicted engagement only on Instagram. By integrating computational methods with self-presentation theory, this study contributes to sports migration scholarship by showing how athletes negotiate hybrid identity and mobilize cross-cultural support in digital sport communication.
- 19 June 2024
- Published by Oxford University Press (OUP) in International Journal of Public Opinion Research
- Vol. 36 (3)
To Live or Not to Live With It? Examining the Cultivation Effect of State Media on Support for the Zero-COVID Policy
This study delved into the influential role of state media in shaping public opinion and cultivating policy support. Focusing on the Chinese government’s insistence on the zero-COVID policy, we conducted a nationwide survey (N = 665) to examine the cultivation effect of state media on Chinese people’s support for the zero-COVID policy. We found that the more state media people were exposed to on social media, the more they supported the zero-COVID policy. State media exposure on social media was positively associated with trust in the government and fear of the virus among the audience, both of which mediated the effect of state media exposure on zero-COVID policy support. By applying the cultivation theory, this study contributed to our understanding of public opinion and information manipulation under authoritarian contexts in the digital age.
- 19 June 2024
- Published by Oxford University Press (OUP) in International Journal of Public Opinion Research
- Vol. 36 (3)
To Live or Not to Live With It? Examining the Cultivation Effect of State Media on Support for the Zero-COVID Policy
This study delved into the influential role of state media in shaping public opinion and cultivating policy support. Focusing on the Chinese government’s insistence on the zero-COVID policy, we conducted a nationwide survey (N = 665) to examine the cultivation effect of state media on Chinese people’s support for the zero-COVID policy. We found that the more state media people were exposed to on social media, the more they supported the zero-COVID policy. State media exposure on social media was positively associated with trust in the government and fear of the virus among the audience, both of which mediated the effect of state media exposure on zero-COVID policy support. By applying the cultivation theory, this study contributed to our understanding of public opinion and information manipulation under authoritarian contexts in the digital age.