Online Echo Chambers: Reinforcement of Anti-Asian and Anti-Immigrant Hate

By: Junyi He Wu ’25

Advising Faculty: Timothy Becker

Major: Computer Science and East Asian Studies

Project/Mission Statement
This project features a cluster of four interconnected Large Language Models, establishing an echo chamber. These Language Models are fine-tuned on anti-Asian and anti- immigrant discourses on mainstream media and social media threads over the course of a week. The cluster reflects the online “echo chamber” that drives misinformation and extremism, as demonstrated by the hate crimes committed toward the Asian populace during the COVID-19 pandemic and the recent legislative oppression of immigrant, transgender, and underrepresented communities. AI-generated portraits of people whose eyes are erased to symbolize the process of Othering and stripping away one's individuality are accompanied by real cases of hate crimes and extremism, which a viewer must step on if they want to interact with the Language Models and be part of the discourse.

Related Fields: Ammerman Center