Created by：Qianlei Ye, Rushi Zhao, Yiwen Chen, Yan Xue, Moxuan Zhang
Overview：The Algorithmic Glass Ceiling investigates how generative AI reproduces occupational stereotypes through image generation. By comparing AI-generated representations of five professions—including CEO, nurse, construction worker, kindergarten teacher and nail technician—the project reveals how algorithms consistently associate particular occupations with specific genders, races and social identities. Through a series of prompt interventions, the group demonstrates that while AI can generate more diverse representations, these often require explicit user guidance rather than emerging naturally. The project argues that these patterns reflect deeper algorithmic structures and training data rather than neutral technological processes. Ultimately, the work encourages audiences to question how AI shapes professional visibility, reinforces existing power relations, and influences our understanding of identity and work in the digital age.
Created with：ChatGPT (DALL·E 3), Gemini, Doubao
Concept
Poster description: The Algorithmic Glass Ceiling explores how AI systems shape our expectations of who belongs in different professions. Through a series of image-generation experiments, the project demonstrates that generative AI often reproduces occupational stereotypes by linking jobs to specific genders, races and identities. Attempts to challenge these defaults reveal both the possibilities and limitations of prompt engineering, highlighting that bias is embedded within the underlying datasets and algorithmic infrastructures rather than individual prompts alone. The work invites audiences to critically reflect on how AI influences representation, visibility and access to positions of power in contemporary society.
Development and Sketches
Contact the Creators
Name: Rushi Zhao
Email: [email protected]