University of Maryland
CASCI

Fall 2018 Talk Schedule

Title: Social Media Impact on Online News Consumption: Echo Chambers, Information Diversity, and Ideological Polarization

February 11th, 2020
11 am – 12 pm
2116 Hornbake Bldg, South Wing

Abstract: There is concern that social media creates echo chambers in which homophily in online social networks and filtering by content selection algorithms combine to limit exposure to diverse viewpoints, thereby encouraging individuals to adopt more extreme ideological positions. Yet, empirical evidence regarding the role of social media in the formation of echo chambers is inconclusive. We demonstrate that this may be due in large part to imprecision and conflation in the definition and measurement of echo chambers. To address this, we provide a more precise conceptualization of echo chambers as narrowed information diversity accompanied by increased ideological polarization. By articulating and separating these two distinct dimensions, elaborating on their potential combinations, and detailing the process by which social media impacts online news consumption, we provide a generative framework for future research. We demonstrate the value of this framework through an analysis of social media usage and online news consumption based on four years of web browsing history for a representative panel of 200,000 US adults. We find that social media usage is associated with a combination of increased polarization and broadened information diversity. These results contradict the prevailing assumption that polarization and narrowing of information diversity are mutually reinforcing and that they typically occur together. Although our findings call in question the existence of echo chambers as typically understood, they nonetheless reinforce concerns about negative impacts of social media on polarization.

The Resilience Approach to Cybersecurity Policy in the Internet of Things Ecosystem

October 2, 2018
11 am – 12 pm
2116 Hornbake Bldg, South Wing

Abstract: The Internet of Things ecosystem includes consumer devices as well as our nation’s critical internet-enabled infrastructure. As we move beyond the threshold of one device per person on the planet, the amount of threat vectors and vulnerabilities is also increasing. However, the ability of policymakers and bureaucrats to respond to cyber threats is limited by status-quo bias and structural impediments to governance such as hierarchy and rigidity. Rather than a top-down policy approach, device insecurity requires solutions from a variety of stakeholders with the goal of achieving resilience in Internet of Things ecosystem. The resilience approach recognizes that the only sustainable way to confront large-scale disturbances is to empower stakeholders at multiple levels to remain persistent in the face of threats.

Speaker: Anne Hobson is a PhD Student in economics at George Mason University and a Program Manager the Mercatus Center at George Mason University. She is also an Internet Law & Policy Foundry Fellow. Prior to her current role, Anne was a Technology Policy Fellow at the R Street Institute and a Public Policy Associate at Facebook. She is an alumna of the Mercatus Center MA Fellowship at George Mason University. She continues to focus on policy issues associated with emerging technologies, such as artificial intelligence, virtual reality, and the internet of things. She received her B.A. in International Studies from Johns Hopkins University.

Making Literature Reviewing Less Painful and More Commonplace: Exploring Sociotechnical Solutions

October 16, 2018
11 am – 12 pm
2116 Hornbake Bldg, South Wing

Abstract: Having a solid grasp of the literature is very often a critical precursor to finding/defining and ultimately solving impactful research problems. But effective literature reviewing is currently painful and often (too) rare. It is very common to hear researchers (both beginning and experienced) struggle with the literature review process, complain about the lack of effective tools for supporting the process, or (if they’ve written a literature review before) express reluctance to do it again. This hampers research progress. My group is exploring sociotechnical solutions to this problem: how do we design configurations of technical systems (e.g., better interfaces, AI) and people (e.g., norms, practices) that can make effective literature reviewing less painful and more commonplace? This is still ongoing work, but I’d love to get feedback on some ways we are framing the problem, and some preliminary data and prototypes we have on hand so far.

Speaker: Joel Chan is an Assistant Professor in the University of Maryland’s College of Information Studies (iSchool), and Human-Computer Interaction Lab (HCIL). Previously, he was a Postdoctoral Research Fellow and Project Scientist in the Human-Computer Interaction Institute at Carnegie Mellon University. He received his PhD in Cognitive Psychology at the University of Pittsburgh. His research and teaching focus on the intersection of people, information, and creativity. He wants to know how they (can best) combine to enable us to design the future(s) we want to live in.

A New Approach for Categorizing Cyber Effects and Measuring their Severity

October 30, 2018
11 am – 12 pm
2116 Hornbake Bldg, South Wing

Abstract: Media is filled with stories about data breaches, phishing attempts, and other forms of cyber activity. Yet these popular accounts struggle to differentiate events which are largely considered a nuisance from those that represent a justifiable public concern. Policy makers struggle with the same dilemma as the inability to discern cyber activity of significance is rooted in the lack of tools that speak to the classification of the effect engineered by the hacker or to its impact on the targeted organization. This talk will focus on a new taxonomy of effect developed at the School of Public Policy, the results of its application to over 2300 published cyber events coded by industrial sector, and new methods to measure the severity of cyber attacks against organizations. The results of this research have direct impact for organizations assessing defensive strategies to combat the problem, policy makers concerned about the diversity of threat to differing parts of the economy, and to the insurance market who seek to offer products that transfer the risk of cyber attacks.

Speaker: Charles Harry is a senior leader, practitioner, and researcher with over 20 years of experience in intelligence and cyber operations. Dr. Harry is the Director of Operations at the Maryland Global Initiative in Cybersecurity (MaGIC), an Associate Research Professor in the School of Public Policy, and a Senior Research Associate at CISSM. Prior to his work with the university, Dr. Harry grew and led a $35 million dollar cybersecurity consulting organization combining analysts and developers to deliver innovative solutions to the private and public sector. His public service includes a 14-year career with the National Security Agency rising to the rank of senior technical leader (DISL). Dr. Harry holds degrees in Economics and History from the University of Colorado, and was awarded a PhD in Policy Studies from the University of Maryland. He is the recipient of the Director of National Intelligence Extraordinary Achievement Medal and the Signal Intelligence Career Achievement Medal. His current research focuses on the development of an analytic framework for assessing cybersecurity risk including the ability to categorize and measure the impact of cyber events.

Documenting Aftermath: An Author-Meets-Critics Session

November 13, 2018
11 am – 12 pm
2119 Hornbake Bldg, South Wing (Note the room change)

Abstract: In this author meets critics session, Drs. Daniel Greene (iSchool), Ricky Punzalan (iSchool), David Tomblin (STS), and Marccus Hendricks (UMD Urban Planning), will discuss Documenting Aftermath: Information Infrastructure in the Wake of Disasters with the author Megan Finn. Documenting Aftermath is an examination of how changing public information infrastructures shaped people’s experience of earthquakes in Northern California in 1868, 1906, and 1989. Finn argues that information orders—complex constellations of institutions, technologies, and practices—influence how we act in, experience, and document events. Examining multiple information orders in different historical moments allows for comparative examination of how people used and rebuilt different public information infrastructures following earthquakes.

Speaker: Megan Finn is an Assistant Professor at University of Washington’s Information School. Her work examines relations amongst institutions, infrastructures, and practices in the production, circulation, and use of information. She examines these themes in a book, called Documenting Aftermath: Information Infrastructures in the Wake of Disasters, with MIT Press coming out in Fall 2018. She completed her PhD in 2012 at UC Berkeley’s School of Information, and spent two  years as a Postdoctoral Researcher at Microsoft Research New England in Cambridge, MA with the Social Media Research Group. Her dissertation research was funded by National Science Foundation Dissertation Grant. Previously, she spent three years working as a engineer and research engineer at HP. I have an Masters degree in Information Management and Systems from UC Berkeley, and a B.S. in Computer Science from the University of Michigan, Ann Arbor LSA Honors College.

Better, Nicer, Clearer, Fairer: A Critical Assessment of the Movement for Ethical Artificial Intelligence and Machine Learning

December 4, 2018
11 am – 12 pm
2116 Hornbake Bldg, South Wing

Abstract: This paper uses frame analysis to examine recent high-profile values statements endorsing ethical design for artificial intelligence and machine learning (AI/ML). Guided by insights from values in design and the sociology of business ethics, we uncover the grounding assumptions and terms of debate that make some conversations about ethical design possible while forestalling alternative visions. Vision statements for ethical AI/ML co-opt the language of some critics, folding them into a limited, technologically deterministic, expert-driven view of what ethical AI/ML means and how it might work. The paper is available here.

Speaker: Daniel Greene is an Assistant Professor of Information Studies at the University of Maryland. He uses qualitative methods to explore the future of work, who is left out of it, and the sociotechnical systems that shape those possible futures. He is currently at work on a book manuscript for the MIT Press tentatively titled The Promise of Access: Hope and Inequality in the Information Economy. Other recent work touches on the organizational history of automated hiring, platforms’ management of freelance mobile developers, the commercial real estate market for internet infrastructure, and the regulation of police body camera data. The paper presented in this talk is a collaboration with Drs. Anna Lauren Hoffman of U Washington’s iSchool and Luke Stark of Microsoft Research Montreal. Daniel lives online at dmgreene.net.

The Fall 2018 CASCI Talk Series is organized by Priya Kumar. Please send questions about the schedule to pkumar12 [at] umd [dot] edu or casci [at] umd [dot] edu.