As the old English proverb has it “the road to hell is paved with good intentions.” Such thoughts spring to mind with the launch of the report Secure Voting by campaigning group WebRoots Democracy. WebRoots are volunteers who ‘campaign for the introduction of online voting in Local and General Elections’. We know where they stand on this issue, but how informed is their argument that online voting can be secure?
Not very informed at all if we are to take their latest report as evidence. The report is essentially an uncritical collage of marketing materials and thoughts from the world of commercial e-voting suppliers. Many of those suppliers are known to ORG as purveyors of systems which we observed going wrong during previous trials in the UK, including Scytl, Everyone Counts and Electoral Reform Services. Problems we observed in those trials included voters unable to cast their votes, Windows having to be re-installed before results could be extracted and errors messages in Spanish as votes were recorded for the wrong candidate.
I don’t disagree with WebRoots Democracy’s desire to boost participation and don’t dispute that low voter turnout challenges the legitimacy of our democratic processes. However the remedy proposed will do little, if anything, to cure the patient. Prime Minister David Cameron understands this and is quoted saying as much:
“Online voting? I mean I don’t have any objection to it, but I think in a way we’re asking the wrong question. The reason people don’t vote is not because it’s too complicated to go down to the polling station; the reason that people don’t vote is because they don’t believe it makes enough of a difference.”
Which is exactly what social scientists and experienced canvassers say too. In short, it is very rare that the logistical difficulty of casting a vote is the reason why someone doesn’t vote. More often the reasons are because a voter feels all parties are the same, their vote doesn’t make a difference, they don’t believe in politics, they live in a safe seat or they don’t feel well enough informed to participate. A technological voting solution won’t solve any of those issues, instead they require the hard work of education and engagement. Which is why, overall, global trials of e-voting have had little or no positive impact on participation rates.
Voting is a uniquely hard problem for computers. Unlike commercial transactions, votes have to be completely secret, anonymous, secure and verifiable. When you shop online it’s not anonymous, banks and shops know who you are and use that to verify your identity. If there’s a problem you can check your statement and ask for a refund. You can’t refund votes and you can’t have a clear voting ‘statement’ to check as that would enable vote-selling and coercion. Which is why a secret ballot is a fundamental right in the UK Human Rights Act, the European Convention on Human Rights and the UN Declaration of Human Rights.
It is computer scientists around the world who have been leading the campaigns against the introduction of electronic voting. These people aren’t luddites, their whole careers depend on the progress of technology. But they understand that a binding political election’s unique properties make them ill-suited to the best computing can offer us today. Thankfully many senior decision-makers are coming to the same view too, which is why after trials countries including Ireland, Italy, France, Germany, Finland and Norway have all withdrawn from the use of e-voting.
I was part of an independent team which studied Estonia’s online voting system and found serious flaws. Our findings, published in a peer-reviewed academic journal, showed that state level attackers could undetectably change the outcome of the elections using Estonia’s online system. And this was despite the Estonian system being the best online system we’d seen in live use, and despite the advantage of every Estonian citizen holding a smart ID card.
The risks of undetectable fraud or error are very significant, the costs of implementing these systems are huge and the benefits marginal at best. So why risk it? At a time when all public services are crying out for investment to go digital there is no compelling case for spending scant resources on e-voting. A report compiling pitches from e-voting suppliers is not going to change the reality that the risks of e-voting are too great for any sound democracy to consider.