Producing Algorithmic Aesthetics: Data Annotation Labour and the Embedding of Human Values in AI
Project dates (estimated):
2024 - 2028
Name of the PhD student:
Xin Bao
Supervisors:
Juli Huang - School of Social and Political Science
Jiazhi Fengjiang - School of Social and Political Science
Project aims:
Xin’s project explores the ethical and social implications of AI production through an ethnographic study of data annotation labour among rural women workers in China. It investigates how everyday practices of image annotation contribute to algorithmic aesthetics and examines how aesthetic standards and human judgments become embedded in AI systems. A central focus of the research is the production of algorithmic aesthetics. Through everyday practices such as labeling, filtering, and evaluating images of human faces and bodies, annotators contribute to constructing what counts as “qualified” data for training GenAI models. These models, in turn, generate images that align with implicit norms of realism, beauty, and acceptability. By de-centering AI as an autonomous system, the project foregrounds the human labour, judgment, and social context through which such aesthetic standards are produced. Xin’ research aims to reveal how the interaction between subjectivitity of human practices and cognition and algorithm influence the process of AI production.
Disciplines and subfields engaged:
Social Anthropology
Science and Technology Studies
Gender Studies
AI Ethics
Research Themes:
Emerging Technology and Human Identity
AI, Automation and Human Wisdom
Emerging Tech and Human Autonomy
Ethics and Politics of Data
Data Justice and Data Violence
Emerging Technology, Health and Flourishing
Emerging Tech and Human Flourishing
Ethics of Algorithms
Algorithmic Justice, Power, Freedom and Equity
Bias and Discrimination in Machine Learning
Algorithmic Accountability and Responsibility
Related Outputs:
Grants and Awards:
2025/26 Tweedie Research Fellowship