My research group has made both constructive and analytical contributions to the study of secure voting systems and statistical election audits. With Filip Zagórski and Neal McBurnett, we are close to releasing a manuscript and corresponding code and experimental data for the Athena class of statistical audits which provide the most efficient approach to ballot polling audits. We have prototyped most of the voting systems we have proposed. Much of our constructive work in secure voting systems is based on collaborations with all or part of the Scantegrity team and Filip Zagórski.
Our work on statistical election audits was used in the audit of the 2020 primary elections in Montgomery County, Ohio. Our work on secure voting systems has been used by the City of Takoma Park in its municipal elections in 2009 and 2011. We have recently worked on influencing policy in the space of election integrity, see my work on Maryland's election procedures, 2016; my testimony for the Wisconsin recounts, Michigan recounts. We have also provided feedback in a number of formats (private emails, open letters, memo to Parliament, invited talk at a conference) to the Indian government on their approach to voting. I currently serve on the Board of Advisers of nonpartisan organization Verified Voting, and on the Coordinating Committee of the nonpartisan Election Verification Network.
In voting system design, our focus has been on the new class of voting system known as end-to-end independently verifiable (E2E) voting systems, where voters and observers can audit an election without being required to trust election officials or voting system software. This property is achieved through the use of a digital audit trail made available by the voting system---on a secure bulletin board such as a website---which can be checked by software written by anyone. Thus, while a certain degree of sophistication is necessary in the software used to perform the checks, access to the information and authorship of the software is not restricted. Our work is distinguished from much other work on E2E-V systems in that our protocols possess the dispute resolution property, where, if a voter observes a problem, she can prove it.
We developed the first non-commercial prototype of an E2E system (Citizen-Verified Voting, WOTE II 2004, of Chaum's ``visual cryptography'' scheme), organized the first workshop on the evaluation of voting systems, VSRW 2006, participated in the development and deployment of (a) the first E2E-V system with ballot privacy ever used in a governmental election, Scantegrity, (b) the first accessible prototype of an E2E-V system, Audiotegrity and (c) the first prototype of an E2E-V remote voting system, Remotegrity.
Our most recent work proposes a remote voting system, Apollo, that addresses the credential stealing attack; note, however, that much work remains before internet voting is secure enough for use in governmental elections.
In 2017, I received the Election Verification Network's Public Engagement Award for my work on Maryland's election audits.
|Current research is sponsored in part by NSF Award 2015253. Past research was sponsored in part by NSF Awards 0505510, 0831149, 0937267, 1137973, 1421373 and the Maryland Procurement Office under contract H98230-14-C-0127.|
Email: my first name at gwu dot edu
Last modified: 14:58:15, Sunday, 21 June, 2020 local time