RAPID: Countering Language Biases in COVID-19 Search Auto-Completes

Computer and Information Science and Engineering - The novel coronavirus (COVID-19) has resulted in sharp increases in online search activity about the disease, its spread, and remedial actions. Hence, search engines can significantly influence public perceptions of the disease and the actions undertaken by the public. If there are language biases in the results of searches, there may also be biases in perceptions and actions taken. This project will systematically analyze the differences in COVID-19 related search auto-completes that are provided to English and Spanish speakers. The results will generate new knowledge on the emergence of algorithmic bias and help ensure equity in health information dissemination at scale amid large-scale health emergencies. The findings will be shared in easy-to-understand terms on an urgent basis in multiple languages to help ensure equal access to health information in the COVID-19 pandemic. This feedback could help improve the health outcomes for numerous individuals facing the COVID-19 pandemic.

The project is designed to yield approaches for countering language-based bias in COVID-19 related health information dissemination by search engines using log analysis and interviews. The first step in this project is to audit the search auto-complete results in Spanish and English, and test if there are systematic differences in the way results are generated across the two languages. The next step is to this utilize focus groups with multiple users to understand how auto-complete queries affect the way English versus Spanish speakers understand COVID-19 disease and take necessary precautions. The findings from the two phases are to be combined to generate guidelines on designing search experiences that support health equity amid a public health crisis. This research topic is likely to attract a diverse range of student researchers, which could help broaden participation in STEM research career pipelines.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.