How technology impacts the democratic process
Experts speak at Elections and Technology symposium
Elections and Technology
Democracy, Citizenship, and Technology Colloquium Series
Bartos Theater
Nov. 1, 2018
A host of issues plaguing the United States’ democratic institutions suddenly and unexpectedly came to the forefront of the country’s political dialogue in the wake of the 2016 national election. Concerns about the vulnerabilities of voting systems and political parties, in addition to the proliferation and dissemination of false information via social media, now plague citizens as they attempt to navigate a rapidly changing political landscape.
These new specters in civic life all have one common thread: they were enabled by the development and of new, often esoteric technologies and the integration of these technologies into the electoral process and into people’s everyday lives. And despite the extensive coverage they receive, their impacts on U.S. democracy and the mechanisms by which they effect those impacts are still unclear, especially to the layman.
To help elucidate the nature of these complex issues, the MIT programs in Anthropology, History, and Science, Technology and Society (STS) invited a multidisciplinary panel of experts specializing in topics relating to the interaction of emerging technologies with society and government to speak about their work to interested attendees. The symposium, entitled Elections and Technology, was held in Bartos Theater and was free and open to the public.
On the panel was Dan Wallach, Professor of the systems group at Rice University’s Department of Computer Science; Daniel Kreiss, Associate Professor and Director of the School of Media and Journalism at the University of North Carolina, Chapel Hill; and Moon Duchin, Associate Professor of Mathematics at Tufts University and Director of the Metric Geometry and Gerrymandering Group (MGGG). In the first hour, each panelist presented for 15 minutes on aspects of their research pertinent to the symposium’s theme.
Dan Wallach spoke first about the vulnerabilities of technologies associated with the voting system. Voter registration databases are connected to the internet, making it possible for hackers to influence them in a partisan way by causing damage. And voting machines, though not connected to the internet, are nonetheless vulnerable to physical tampering. Wallach shared an amusing, though troubling anecdote in which Defcon participants were able to easily manipulate decommissioned voting machines at the annual computer hacking convention.
Wallach concluded by describing several voting and vote-counting procedures that could be adopted to improve the security of voting systems. Casting votes on paper ballots would eliminate the vulnerabilities associated with voting machines, while machines could be used to scan the papers and tally the votes to enable swift and accurate counting. Then, to establish confidence that the count is correct, risk-limiting audits could be performed in which a random subset of the electronic tallies would be double-checked with the corresponding paper ballots and a margin of error in the count would then be established. Additionally, advanced cryptographic techniques could be employed to give voters a “voting receipt” that they could use to verify with a third party that their vote was counted correctly without revealing for whom their vote was cast.
Moon Duchin spoke next about her research on mathematically modeling congressional redistricting in the U.S. and using these models to identify gerrymanders. She first framed the problem mathematically, illustrating the immense combinatorial space that arises in trying to divide Pennsylvania into simple polygons with roughly equal populations. For all cases, with many simplifying assumptions, the number of valid combinations was in the trillions. She then laid out the legal guidelines for redistricting, namely that districts must have roughly equal populations, be contiguous, and respect the Voting Rights Act of 1965 (i.e., that they must be drawn in such a way as to ensure minority communities receive political representation). “I’m doing a mathematical model… and trying to make contact with political and cultural realities in very complicated ways,” she said.
Duchin described a “random walk” approach used by her group to “sample from the ginormous space of possible plans” rather than enumerating them all explicitly. “In not too long… I can find the reasonable range of properties I care about,” she said. Duchin then showed how her group had applied this approach to analyze the electoral map of Virginia, which recently had its House of Delegates map thrown out as a racial gerrymander by a panel of federal judges. She showed a plot with the districts arranged from “least black” to “most black” on the x-axis and the number of black constituents on the y-axis, She then superimposed the Democrat and Republican redistricting plans on the graph. She demonstrated that the Democrats’ districts tended to cluster higher on the y-axis, where it was more likely that districts would be able to elect black representatives. The Republican plans tended to cluster nearer to the center of the y-axis, where the effect of the black vote was greatly reduced. “I hope that what we’ve done is created a tool that gives you the ability to recognize unstated design principles that went into making a map,” Duchin concluded.
Daniel Kreiss spoke last about how technology is shaping the behavior of political actors (e.g. candidates, campaigns, political parties, etc.). He began by describing three eras of civic engagement in the United States, proposed by political scientist Pippa Norris, based on the role technology played in political dialogue and campaign advertising. The first is the Labor-Intensive Era, in which political organizations were very active in directly contacting and mobilizing their constituencies and in which voter turnout was extremely high. The second, the Capital-Intensive Era, was characterized by large investments in broadcast advertising, reliance on public opinion polling and more passive modes of civic participation. In the third era, the Postmodern Era, campaigns and candidates struggled to understand and properly utilize emerging technologies, epitomized by Howard Dean’s infamous scream that spelled doom for his 2004 campaign after going viral.
He then proposed a new era — the Technology-Intensive Era — based on his own research that, in conjunction with the Postmodern Era, characterizes the present-day. “In the Technology-Intensive Era, everything campaigns do now has an underlying technological, and often data component to it,” Kreiss said. Using that data, political campaigns attempt to maximize potential sympathetic voters’ exposure to their message. “The same message that might be coming to me through a sponsored tweet is being reinforced at my door, is being reinforced in my mailbox… is being reinforced by a canvasser that’s knocking at my door.” This new paradigm has given rise to a vast ecosystem of firms specializing in tech and digital data analytics that provide expertise to political parties and campaigns. In this way, the professional landscape of political campaigns has been fundamentally altered, Kreiss explained.
The Technology-Intensive Era has also resulted in the emergence of platform companies, such as Facebook and Google, as important political actors. “ [These] platforms are now significant shapers of the political communication advertising infrastructure,” Kreiss said. And because “likes” and “shares” enhance the exposure of posts on these platforms, political actors are now compelled to generate emotionally inflammatory content that is more likely to grab the attention of users. Kreiss concluded by suggesting that this development has led to the emergence of an era of “clickbait” advertising in political campaigning.
Though all the panelists spoke about fairly disparate topics, the common theme was evident: technology plays a pivotal role in shaping democratic institutions and how citizens engage with them. In some cases, technology introduces new problems to the electoral process, such as the vulnerabilities of electronic voting systems and the polarizing nature of platform-based political dialogue. But technology can also provide opportunities to address previously intractable problems, exemplified by Moon Duchin’s work on mathematically modeling gerrymandering. This duality highlights the importance of the panelists’ research: by elucidating technology’s influence on the political process, we can better understand how to use it in ways that maintain the health of our democracy.