Next week we continue with the series of seminars on Security & Privacy at INESC-ID. If you're interested in the activities of the Security & Privacy Thematic, join at: https://mattermost.inesc-id.pt/signup_user_complete/?id=eptffb51ii81mqrak34m991roy&sbr=su

Seminar: From Individual Verifiability to Universally Verified Cast as Intended

Professor Peter Ryan
University of Luxembourg
https://www.uni.lu/en/people/peter-y-a-ryan/

Abstract:
End-to-end verifiability has long been the holy grail in the quest for demonstrably trustworthy elections. The goal is to ensure that election systems generate sufficient evidence in the course of an election to allow voters and observers to detect any corruption of votes.
Many such schemes have been proposed, and these typically involve certain types of checks that can only be performed by the voters, and not by other observers. These so-called "individual" audits include ballot audits, cast-as-intended, and checking the presence of the ballot in the bulletin board, recorded as cast.

Making voters a foundation for the trust, rather than having to trust officials, vendors, code etc., is of course laudable. However, individual audits place a burden on voters that is hard to motivate. Most voters prefer simply to vote and go, and, in practice, few perform any checks. The latter is an issue because if too few voters perform such checks with sufficient diligence then the election cannot be deemed to be verified, even if it is run on a verifiable system.

Here, we describe constructions that allow individual audits to be outsourced to independent auditors and observers, transforming individual checks to being universal. There is no longer any need or expectation for voters perform ballot audits (but the option is still available to them).
This contributes to improved usability and acceptability. It ensures that elections will be fully verified, not simple verifiable.

Ballot audits take the form of verifying cryptographic signatures rather than verifying encryptions, avoiding the need to reveal randomization factors, etc. For transparency and usability, the ballots display a clear text version of the vote to the voter.  Dispute resolution is improved as ballot auditing is simply signature verification.

Key References:
1- MOSAHEB, R., ROENNE, P., RYAN, P. Y. A., & Sarfaraz, S. (2025). Direct and Transparent Voter Verification with Everlasting Receipt-Freeness. In D. Duenas-Cid (Ed.), Electronic Voting - 9th International Joint Conference, E-Vote-ID 2024, Proceedings. Springer Science and Business Media Deutschland GmbH. doi:10.1007/978-3-031-72244-8_8
https://hdl.handle.net/10993/63432
2- Damodaran, A., Rastikian, S., ROENNE, P., & RYAN, P. Y. A. (2025). Hyperion: Transparent End-to-End Verifiable Voting with Coercion Mitigation. In V. Nicomette (Ed.), Computer Security - ESORICS 2025 - 30th European Symposium on Research in Computer Security, 2025, Proceedings. Springer Science and Business Media Deutschland GmbH. doi:10.1007/978-3-032-07891-9_7
https://hdl.handle.net/10993/66956
3- RYAN, P., Demirel, D., Henning, M., van de Graaf, J., & Buchmann, J. (2013). Prêt à Voter providing everlasting privacy. Lecture Notes in Computer Science, 165-220. doi:10.1007/978-3-642-39185-9_10 https://hdl.handle.net/10993/25426

When: Thursday, January 8th, 3pm

Where:  INESC-ID Alves Redol, Auditorium (room 9)