AI Opaqueness: What Makes AI Systems More Transparent? (Panel)

Abstract

Artificially Intelligent (AI) systems are pervasive, but poorly understood by their users and, at times, developers. It is often unclear how and why certain algorithms make choices, predictions, or conclusions. What does AI transparency mean? What explanations do AI system users desire? This panel discusses AI opaqueness with examples in applied context such as natural language processing, people categorization, judicial decision explanations, and system recommendations. We offer insights from interviews with AI system users about their perceptions and developers’ lessons learned. What steps should be taken towards AI transparency and accountability for its decisions?

Victoria L. Rubin

Faculty of Information & Media Studies, Western University

Victoria L. Rubin is Associate Professor at the Faculty of Information and Media Studies (FIMS) at the University of Western Ontario, the Director of the Language and Information Research Lab (LiT.RL). She broadly specializes in information retrieval and natural language processing techniques, currently working on automated detection of misinformation and disinformation in news.

Jacquelyn Burkell

Faculty of Information & Media Studies, Western University

Jacquie Burkell is (Acting) Associate Vice-President (Research) and is an associate professor in the Faculty of Information & Media Studies. She holds a PhD in Psychology (Cognitive Science) from Western and Jacquelyn served as the faculty’s Assistant Dean of Research for seven years and chaired the Associate Deans (Research) group from 2016-2018. Throughout her career, Jacquelyn has served on a wide variety of academic committees, including the 2016 URB Task Force Steering Committee – Support for Research in Social Sciences, Arts, and Humanities at Western. A highly collaborative scholar, Jacquelyn is a co-investigator on two SSHRC partnership grants – one examining artificial intelligence in the context of justice, the other focused on youth equality and privacy online. More broadly, her research focuses on the social impact of technology and examines how technological mediation changes social interaction and information behaviour.

Sarah E. Cornwell

Faculty of Information & Media Studies, Western University

Sarah E. Cornwell is a doctoral candidate in the LIS program at FIMS. Building on previous degrees in linguistics (MA), and Anthropology & Cognitive Psychology (BAS), her research interests include multilingualism, natural language processing, and everyday information seeking. In essence, she focuses on the interaction of information technologies and human linguistic diversity.

Toluwase Asubiaro

Faculty of Information & Media Studies, Western University

Toluwase Asubiaro is a PhD candidate at the Western University’s LIS program. He holds a B.Sc. in Mathematics and Master of Information of Science. He started his career in Language Technology in 2012 as a volunteer research assistant in African Languages Technology Initiative (ALT-I), Ibadan, Nigeria. His major research interest is informetrics, natural language processing, information retrieval and automatic language identification. Prior to his present position in LiT.RL, he had contributed to studies on language technology for Yoruba, a Nigerian language.

Yimin Chen

Faculty of Information & Media Studies, Western Univeristy

Yimin Chen is a PhD candidate in Library and Information Science in the Faculty of Information and Media Studies at the University of Western Ontario. His research examines the communicative practices of online communities and cultures, with a focus on Internet trolling behaviors and the controversy surrounding them. His previous projects range from fake news and deception detection, to library automation, to the impact of political memes on social media.

Danica Potts

Faculty of Information & Media Studies, Western University

Danica Potts is a PhD student in LIS at Western University interested in the social and ethical issues of artificial intelligence. Her PhD research focuses on the nature of trust between humans and artificial intelligence and the implications of trust on users’ information behaviour and decision-making processes. Other projects Danica is a part of look at algorithmic decision making and explanations, and algorithmic literacy.

Chris Brogly

Faculty of Information & Media Studies, Western University

Chris Brogly is a Doctoral Student in the Health Information Science program at the University of Western Ontario. He completed his MSc in computer science from Western in 2017 and a BCS Co-op from the University of Windsor in 2015. His research interests are in health informatics, predictive modelling, and the internet.