Censorship’s Implications for Artificial Intelligence
In this webinar, Jeffrey and Molly will discuss a recent paper Molly co-authored with Eddie Yang. Abstract: While artificial intelligence provides the backbone for many tools people use around the world, recent work has brought attention to the potential biases that may be baked into these algorithms. While most work in this area has focused on the ways in which these tools can exacerbate existing inequalities and discrimination, we bring to light another way in which algorithmic decision making may be affected by institutional and societal forces. We study how censorship has affected the development of Wikipedia corpuses, which are in turn regularly used as training data that provide inputs to NLP algorithms. We show that word embeddings trained on the regularly censored Baidu Baike have very different associations between adjectives and a range of concepts about democracy, freedom, collective action, equality, and people and historical events in China than its uncensored counterpart Chinese language Wikipedia. We examine the origins of these discrepancies using surveys from mainland China and their implications by examining their use in downstream AI applications.
Date: 28 October 2020, 17:00 (Wednesday, 3rd week, Michaelmas 2020)
Venue: https://www.crowdcast.io/e/Roberts&Ding
Speakers: Jeffrey Ding (University of Oxford), Margaret Roberts (UC San Diego)
Organising department: Faculty of Philosophy
Organiser contact email address: fhiadminassistant@philosophy.ox.ac.uk
Part of: Governance of AI Seminar Series
Booking required?: Required
Booking url: https://www.crowdcast.io/e/Roberts&Ding
Cost: FREE
Audience: Public
Editor: Anne Iwashita-le Roux