Censorship of Online Encyclopedias - Implications for NLP Models

Explore the impact of online encyclopedia censorship on natural language processing models in this 18-minute conference talk from the FAccT 2021 virtual event. Delve into research by E. Yang and M. Roberts examining how political censorship of online information sources affects the development and performance of NLP systems. Gain insights into the potential biases and limitations introduced when training data is restricted or altered due to censorship practices. Understand the broader implications for AI fairness, accountability, and transparency in contexts where information access is controlled or limited.