講題：Nudging Users towards Exposure Diversity: How can News Recommender Systems Contribute to Well-functioning Democracies?
講者：Nicolas Mattis (Visiting PhD Student, Communication Science from Vrije Universiteit Amsterdam)
時間：2021 / 2 / 18（週四）下午 2 點
地點：中研院 人社中心 第一會議室
There is a broad consensus that citizens’ exposure to a diverse range of topics and viewpoints is necessary, or at the very least beneficial to functioning democracies. As (personalised) recommendation algorithms become increasingly important in determining individual news diets, this ideal of exposure diversity is often said to be under threat. However, recommender systems and their underlying algorithms do not have to trap individuals in filter bubbles or echo-chambers by default. They can also be deliberately designed to contribute to healthy democracies by facilitating exposure to diverse information.
Situated in a broader interdisciplinary project ( at Vrije Universiteit Amsterdam & University of Amsterdam) that examines how news recommender systems (NRSs) can and should contribute to democracy, this PhD project aims at a) modelling individual news selection in NRSs and b) exploring if, how, and to what extent their design can facilitate exposure diversity and thus contribute to pro-social outcomes.
Against the normative background of the overall research project, this talk will explore core considerations and strategies for diversity-facilitating NRS design from a communication science perspective. It will cover potential opportunities and pitfalls of diversity-facilitating NRS design and lay out a future research agenda for user-recommender interactions and their effects.