The Computer Science (CS) community has witnessed decades of research into security techniques for commonplace transactions such as banking or voting. Yet, powerful cryptographic techniques - such as encrypted ballots for voting - have not made their way into the public domain. A large part of the reason is that the current standard for electronic voting is pass/fail in nature, and provides no incentive for the use of novel techniques. Further, while desired properties of voting systems (such as count integrity and ballot secrecy) are mentioned in the standard, they are not well-defined, nor is the standard centered on the properties. Instead, the standard is centered on specific designs of voting systems from over fifteen years ago. As a result, innovative systems, in addition to having their advantages obscured by a pass/fail standard, also face hurdles for not being based on the old designs.
This project is aimed at pursuing an important near-term (Spring 2005) opportunity to change the basis on which electronic voting systems are evaluated. Technical recommendations are to be made to the U.S. House of Representatives in Spring 2005 by the Senate-appointed Election Assistance Commission (EAC), under the Help America Vote Act (HAVA). This group of researchers hope to contribute the recommendations. In particular, the group is working together to rate and compare voting systems on performance with respect to well-defined properties such as: ballot secrecy, count integrity, voter and public verifiability, system transparency, reliability, usability and accessibility for the handicapped. The requirement for such a framework is urgent, as the HAVA deadline is firm and another opportunity to contribute to the conduct of mass elections in the U.S. may not appear again soon.
This project will contribute to the technical core of the recommendations to the Senate by working closely with the Technical Guidelines Development Committee (TGDC) of the EAC and the Voting Systems Standards Project of NIST. The project will develop a rating mechanism that captures the properties of the flow of information among voters, voting machines, aggregation processes, counting and auditing. The investigators hope to test a prototype against some of the key dimensions of the performance rating system developed. The project will also develop a test suite to enable testing with respect to some of the properties for large numbers of votes.
Intellectual merit: The project will help refine the debate on voting systems by defining a technical core for ratings of performance with respect to specific properties. It will also develop a test suite and study the scalability of a leading cryptographic technique. Outcomes of this project will include (a) the security and privacy aspects of the technical core used to evaluate voting systems and (b) an open-source prototype voting system.
Broader impact: The project will contribute to the process of determining a new voting standard through input to the technical recommendations of the EAC and will make available to public interest groups for analysis and education the first open-source prototype of a voting system with voter verifiability. The project will also help train graduate students in the area of security, and will help develop ties between the computer science community and public interest groups engaged in democratic processes.
