Voting from a computer or smartphone sounds certainly convenient and fascinating. An increase in turnout and the comfort of working from home is a strong incentive for its strong development in the future.
However, studies have been showing so far the opposite outcome: since it is still seen with suspiciousness, this leads to less turnout in practice and a major disenfranchisement of the public.
Just for the fact that the public understanding of electronic voting systems perceives this as vulnerable to many failures reduces, in fact, an election’s legitimacy, with consequences for democracy. But there is a lot more.
Sven Heiberg, Product Manager at Smartmatic-Cybernetica Centre of Excellence for Internet Voting OÜ, presented his concerns during Paris Blockchain Week Summit 2020.
Sven leads the development of the online voting platform TIVI and has worked in the development of the Estonian online voting system since 2004.
Saving integrity and confidentiality is central to the debate.
Can we relax the tension between public verifiability and confidentiality?
In popular use, online voting and Internet voting are interchangeable, meaning any system allowing voters to express votes via the Internet (including Blockchain-based systems), whereas electronic voting refers to any system where votes are cast purely electronically.
Voting online seems naturally easy at first sight: just a few steps to take on a phone from a remote location, without having to cue outside a polling station physically. But fatal flows may occur.
In paper-based voting methods, transparency is achieved through physical observation of procedure, and even these are not totally immune to serious failures.
If hackers exploit a computer’s hardware or software such as a targeted voting system, they take full control of the way this also communicates with the voters.
Human incapability to observe electronic processes suggests that an independent audit of the data (created by the participants of the voting protocol) is needed.
First of all, ballot secrecy is a top priority to prevent corruption and coercion.
Software independence is another essential requirement to ensure monitorability of the casting and counting stages of an election so that an occurred error in a system’s software is detectable.
This means that a system for ballot casting must produce an evidence trace with a linked verification procedure to check that the system recorded votes as intended and counted them as collected.
The one-million-dollar question “who should detect the errors?” is still answerless but surely individual voters should be able to verify that their ballot reflects their intended choices.
Elections should also be audited in terms of compliance audits and risk-limiting audits: we need to ensure that each relevant part of the system is functioning correctly as intended to guarantee that the electorate could legitimately be assured that the elections found the real winner.
To sum up, software-based systems are much more exposed to failure than non-software-based systems.
Experiences around the globe
It is commonly said that online voting can improve the turnout being also cost-effective.
Studies have so far not proved any positive impact on turnout, for instance, in Switzerland or Belgium, or not a very significant improvement as in Canada.
E-voting could indeed lower the cost of participation for some population: overseas voters, people with disabilities, or the youngest generations, generally between 18 and 25 years old, usually distant from the political process.
Recently, Oleksandr Stelmakh, a member of the Ukrainian Central Election Commission (CEC) organised an e-voting test to examine the use of Blockchain during elections, given the impossibility of modifying the information stored in the Blockchain. However, he then mentioned that a node in each police station would be required to vote, with a cost of 1000 euro each.
An experiment of Blockchain-based voting system was also used in Russia, for the city council elections of 2019: the test revealed a severe vulnerability of the system, despite the presence of security researchers invited to audit and a decent amount of transparency.
Estonian elections showed that e-voting might favour higher-income and higher-education demographics to participate in the ballot.
Sven Heiberg and his colleagues are currently working on a voting protocol that allows publishing cryptographic trails necessary for the integrity verification in a way that combines the reliability of the process without sacrificing the ballot secrecy.
The risk of rushing into a quick “high-tech” solution is to give not enough attention and time to independent audits and security experts.
Compromising the democracy process is a very high price to pay and it might suggest that there is still a long way to come.
E-Voting, a new phenomenon?
E-Voting, a new phenomenon?
In December 2018, NEM a dual-layer blockchain with a NEM Infrastructure Server launched the e-voting initiative in Ukraine.
Watch the video below to find out more: