I have been working in the IT field since 1979; So I saw a lot of technologies pass by Among these technologies, there are, as everywhere, fashions. The fashion lately has been for " blockchains ", which are said to solve everything from dysentery to returning late library books to clap.
For the past twenty years, I have also been working on elections; most recently, I worked on the team planning municipal elections. And I, of course, worked in other elections, federal, school, provincial, where I held quite a lot of positions.
In the case that interests us, namely elections by Internet (and therefore computerized), we want to apply to a given field technologies which have amply proven themselves without asking what is the real goal sought.
In the case of banking transactions, for example, we need :
- confidentiality
(so that it remains between the bank and the customer), - positive identification
(so that it concerns ONLY the bank and the customer - just as much as the bank. needs to know that the customer is really the customer and not a scammer, the customer needs to know that the bank is the bank and not a phisher) - traceability
(in case of problem or simply audit, without taking into account what is required by the regulations).
All these conditions are reasonably fulfilled by the means currently implemented, to the point where the banks do everything to discourage customers from having recourse to a cashier in the flesh...
But in the event of an election, what are the needs? They are not at all the same as those of a bank.
We also need
- confidentiality
(because the vote is secret), - positive identification
- (to avoid fraudulent votes),
- but no traceability
(because the vote is secret).
On the contrary, the vote cast must cease to be traceable from the moment it is placed in the ballot box. In addition, there must be a certain degree of transparency (which stops, of course, with the secrecy of the vote) to guarantee to any observer that the voters who vote are indeed entitled to vote, that their vote is indeed counted. and that it remains completely secret.
A proven system
These are conditions that conflict with others, but with which the current system has absolutely no problem with. As a reminder:
- once the legitimate voter has been identified,
- he is given a ballot paper initialed by the deputy returning officer and provided with a numbered detachable stub;
- once this ballot is filled out, the correspondence of the number on the stub with that of the stub remaining on the notebook ensures that it is indeed the ballot given to the voter and not a "telegram" ;
- the stub is then removed before putting the ballot in the ballot box, which then stops the traceability of the ballot and thus guarantees the secrecy of the vote.
Better still, paper operations can be followed and monitored by absolutely anyone with a minimum of explanations, unlike processes implemented on a computer which is a black box whose operation will remain opaque, even to the greatest specialists, for the simple reason that it is impossible to follow in real time the functioning of a computer simply because of its speed.
Public confidence is absolutely essential in the conduct of the poll, because it is the survival of our democratic system. The fact that the current system is completely traceable by ordinary people matters a lot. The introduction of a black box at any point in the processing of votes breaks traceability. And what about the trendy blockchains , which are far from obvious how they work (I think I understand how they work, but I'm not at all sure - at least, not sure enough to trust them personally, despite my years of experience).
With electronic voting, we can never have absolute certainty that a failure (intended or not) in the system will not hamper the accuracy of the results. How to guarantee that an electronic voting system does not present inaccurate results, not decided in advance by I do not know who can afford to "afford" a puppet government?
Why change?
And why would we want to computerize the vote? To go faster ?
It takes between one and two hours to count the votes for each ballot box. In four hours, you usually have 99.99% of the results countrywide.
To save money? Of course, the current system needs a lot of staff. But 99% of the staff only work on election day. And we have a poll every two years or so. It's not like we have an election every week ...
And here, I'm only talking about the staff; What about the expense of developing a safe and reliable system that would not risk undermining public confidence?
To increase the participation rate? As they say, “the absent are always wrong”. It is true that the low turnout can be alarming. But it is far from guaranteed that the introduction of gadgets will attract more people; other countries have simply adressed the problem by making voting compulsory.
No, before making the immense effort to implement Internet voting, we could make voting compulsory, as in Belgium and Australia.
Electronic voting is a “solution” for the tech-savvy looking for a problem that can be solved much better otherwise.