Supporting Reader-Oriented News Experience for Immigrants through Chatbot Design

Abstract

News reading helps individuals stay informed about societal developments. For immigrants, local news serves as a crucial information source for adapting to their host country. However, prior studies have highlighted that challenges, such as language barriers and a lack of relevant social knowledge, often hinder immigrants’ ability to fully comprehend and utilize news content in their everyday lives. With the digital transformation of media, there is a growing interest in enabling readers to interact with news content based on their needs and interests. Emerging technologies such as chatbots hold promise for supporting this type of engagement through intuitive interactions. As recent advancements in large language models have further unlocked the potential of chatbots, it is important to consider what needs immigrants have when reading news with the assistance of a chatbot, and how chatbots can be designed to provide a reader-oriented news experience.This dissertation seeks to address these gaps through a multi-study project. The first study investigates immigrants’ news reading needs using a between-subjects experiment involving Chinese immigrants and U.S.-born locals, where participants read a set of news articles with the assistance of an LLM-powered chatbot. This work presents a typology of reader-initiated questions and reveals differences in how these groups engage in analytical thinking about news content and seek practical guidance from it. The second study extends this research by including a Vietnamese immigrant group, offering a cross-validation of the findings from the first study. Study Two enriches our understanding of reading needs across different immigrant groups. My third study aims to extend both the empirical understanding and the design thinking for immigrants’ news experiences by connecting their voices with those of journalists. I invited 11 immigrant readers and 7 journalists from the metropolitan area in the U.S. capital region into a co-design study. Participants joined a series of interlinked activities spanning multiple days, and they collaboratively envisioned technology design for immigrant readers by situating their design speculations within existing links between news production and consumption. The findings of this research surfaced multiple sets of (mis)alignments between immigrant readers and journalists on values that should be prioritized in news topic selection, content presentation, and information processing. Grounded in these insights, participants articulated four metaphors for future news technologies serving immigrant readers. The findings contribute meaningful insights for enhancing immigrant readers’ news experiences through the appropriate alignment of roles among technology, journalists, and immigrant readers. These findings extend the first two studies by contextualizing the identified news reading needs within immigrants’ lived experiences and real-world journalistic practices, and by generating design insights for technologies that better support immigrant readers. This dissertation contributes: 1) an enhanced understanding of immigrant news readers’ needs through empirical findings, such as a typology of reader-initiated questions serving different purposes; and 2) actionable insights for designing chatbots that deliver audience-oriented news experiences, grounded in collective input from stakeholders in the news ecosystem.