Column

Internet Voting Revisited

The Chief Electoral Officer (CEO) of Ontario recently published a report on ‘alternative voting technologies’, mainly ‘network voting’ by Internet or telephone. The Election Act had been amended in 2010 to require such a study. His conclusion, somewhat controversial, is that no election technology currently exists that can satisfy the criteria that his team decided should apply to any such system before it could be implemented.

The stated criteria are a helpful base line for considering the use of voting technologies. No authoritative list could be found, so after study, the list was created. Most people who have commented have accepted the criteria, even if they challenge the conclusions.

  • accessibility – including to voters with disabilities
  • individual verifiability – so the voter can be sure his or her vote has been properly deposited
  • one vote per voter – though multiple votes would be OK if only the last vote counted
  • voter authentication and authorization – identity and eligibility to vote
  • count votes only from eligible voters
  • voter privacy – no one can connect a voter with his or her vote, at any stage of the process
  • results validation – election officials can verify that the result represents the will of participating voters
  • service availability – the process and its components (voters’ list, etc.) will be available to everyone with an interest

A question not provided for in the criteria but which is worth debating is whether the voting system should be based on open-source or proprietary software. It is arguably easier to check an open-source system for security flaws, but it is easier to find and acquire the rights to use proprietary systems.

The report goes on to list the benefits and the risks of network voting. Not a lot of surprises in the lists, except possibly the risk that a network system could be very costly, especially when supplementing existing voting channels. The cost of setting up a pilot project in a by-election is mentioned later (~$2 million and more than six months’ planning).

The ‘potential benefits’ are said to be these:

  • increased choice for electors in how and when they cast their ballot
    • voting is more accessible to electors with disabilities
    • voting is more accessible to electors outside the jurisdiction, such as military personnel, snowbirds and students
  • fast and accurate tabulation of votes is possible, which is particularly useful for complex ballots with multiple choices and maybe even a referendum
  • easier preparation and voting in places with frequent elections
  • reductions in the number of election workers and locations
  • cost reductions ‘may be achieved’, especially if paper ballots are eliminated entirely
  • environmental benefits from less travel for officials and the use of less paper.

The ‘risks and limitations’ that would need to be mitigated are stated to be:

  • security breaches that could jeopardize the integrity of the voting process
  • unavailability of secure digital authentication mechanisms
  • possibility of deliberate or inadvertent denial of service
  • lack of transparency, including on a recount or audit, because of lack of paper trail
  • the digital divide – some people do not have equal access to the Internet (though phone voting might mitigate that at an early stage)
  • ‘network voting is costly’.

The report reviews other Canadian studies, including a British Columbia discussion paper from 2011 and a technical report done for Elections Canada in 2010, and mentions some others on Internet voting. It examines in detail the experience in Canadian municipalities in this area, as well as that in a number of other countries. It is a thorough review and provides a useful bibliography in the field as well.

The result in sum is that very few jurisdictions are actually engaged in network voting beyond the municipal level, except in Estonia, where the system relies on ubiquitous government-issued electronically-capable identity cards. Several wider-scale projects have failed, some dramatically (a U.S. effort to provide remote voting for the military cost over $60 million and attracted 84 votes.) Even municipalities have often had trouble with their equipment, needing to extend voting hours or even adding a day. There is considerable reluctance by those responsible for making elections happen in a reliable way to expose their work to such hazards.

The 2010 amendments to the Ontario statute also allowed the Chief Electoral Officer to run pilot projects in by-elections using alternative voting technologies. Expanding any such pilot to the provincial scale was subjected to a number of strict conditions, including public hearings in the Legislature.

Early in the review process, Elections Ontario planned to run such a pilot, probably in 2012. The new report includes a Business Case for setting and operating such a project, published in 2011. However, once the criteria for implementation were established and suitable technology was found to be lacking, the idea was dropped (the conception having allowed a number of ‘off-ramps’ if things weren’t going well.) The Business Case is still a useful compendium of the demands of operating such a system.

The CEO has no current plans to run any such pilots, though he recently told the Toronto Star that one might be possible in 2017 (which the Star’s headline interpreted as ‘Elections Ontario says it’s time for a pilot project’).

It is interesting to note that Elections Canada has come to similar conclusions. In a report published in the spring of 2013, it said this (at page 35):

In recent years, Elections Canada studied the ways and means to provide Internet-based voting. At one point it publicly stated its intent, contingent on Parliamentarians’ approval, to pilot I-voting in a by-election after 2013.

However, during the summer of 2012 Elections Canada’s senior management decided to scale back efforts on Internet voting and delay any I-voting pilot project until after the next general election. Reasons given for this decision included high costs and risks during a period of fiscal restraint, a critical lack of publicly available user “authentication” methods, and the fact that expected benefits and gains in efficiency are likely several years away. Modest research efforts are being continued.

Current Internet voting systems carry with them serious, valid concerns about system security, user authentication, adequate procedural transparency, and preserving the secrecy of the vote.

The decision of the Ontario CEO to postpone implementation of network voting (while continuing to review progress in technology and techniques) has come in for some criticism. The main source of the criticism is advocates for people with disabilities, who see technological voting as an important way to improve access to the voting process. A number of amendments made to the Act in 2010 responded to submissions from this group, including the requirement for the current Report.

Thus the Accessibility for Ontarians with Disabilities Act Alliance (AODA Alliance) called the CEO’s report ‘a slap in the face’ for its community. ‘[R]eading it more carefully, this Report is a formula for more delay and foot-dragging.’ The Alliance says that the Election Act ‘imposes an unjustified total ban on telephone and internet voting, technology that can make the voting process truly accessible. It gives the power to lift that arbitrary legal ban to Ontario’s unelected and unaccountable Chief Electoral Officer.’ (The CEO is an officer of the Legislative Assembly and reports to the Speaker. He is appointed on a joint address of the Legislature and can be removed the same way.)

The Alliance gives short shrift to the CEO’s expressed concerns about the implementation criteria for network voting, mentioned at the outset of this note. It points out that Ontario accepts mail-in ballots, and they are no more secure than the network voting done by municipalities. This is, it says, a double standard on the security question that Elections Ontario puts at the centre of the discussion (instead of accessibility, to which the Alliance would give priority.)

It refers to its ‘main contention’ that if we can reliably bank by phone or online, and file our taxes online, ‘why can’t the Government figure out how to securely vote online or by phone?’ The apparent parallel between online banking and electronic voting was explored in an earlier Slaw column.

Another advocate of online voting is the Ontario Human Rights Commission. In a submission to Elections Ontario late last year (the Commission does not appear to have formally commented on the 2013 report), it accepted the implementation criteria but urged adoption of accessible voting by phone or Internet. The Commission wanted to see a pilot done in an early by-election, or even in a general election, subject to special limits and focus.

The Commission also objected to the ban on using voting machines on the day of the election itself, their use being restricted to advance polls. A similar precaution was urged in the recent report, and in fact exists in several municipal online voting systems as well. The principle of the restriction is that it gives voters the chance to vote in person if the machines fail. If they fail on election day, providing a remedy may be difficult.

It can at least be said that the debate is not only squarely joined, but better resourced as a result of the recent report – which, as noted, draws in work at the federal level and BC, as well as the municipal and international experience. (My earlier article refers to some other sources, including computer security experts, who are very sceptical of remote voting of any kind.) The Toronto Star has come out in favour of at least a pilot project (reading the CEO’s report as much more positive on that prospect than does the AODA Alliance), but a member of the Ontario Legislature has expressed his doubts in the pages of the same paper.

While Elections Ontario and Elections Canada continue informal review, without express plans or timetables, British Columbia has established an independent Internet Voting Panel to conduct a study for that province, building on the 2011 paper. Meeting notes of its early sessions are online, and a preliminary report is expected in October 2013.

All of the official publications express confidence that the barriers will be removed, and that technology will eventually be produced that satisfies the implementation criteria or other concerns expressed about remote voting. The debate turns on timing and priorities as well as on the technology. As usual, decisions to proceed will depend on political as well as technical considerations, and in an area as central to our democracy as a legitimate voting process, that is no doubt as it should be.

Retweet information »

Comments

  1. David Collier-Brown

    Since the discussion of e-banking in http://www.slaw.ca/2010/10/25/electronic-voting-and-the-law-its-not-like-e-banking/
    I’ve been the victim of ATM fraud and found that my bank (and others) are spending a very significant amount of money reimbursing people who were defrauded courtesy of the bank.

    I was doubtful of e-voting security before: now I’m reasonably sure it’s a chimera. Probably composed of a pig, a dodo and a skunk (;-))

    –dave

  2. Well written summary! I would contend that all these reports need to be sufficiently neutral, and not necessarily advocate for internet voting at this point, but for government to proceed with caution. This is because the world of Internet Voting provisioning is still wildly divergent in terms of security, privacy, accessibility, auditability and so on. Many companies that CLAIM to offer internet voting are really just hopped up survey tools without the necessary checks and balances in the background to ensure the voter’s choices are kept in line with election standards. And, in these cases, may the contracting office beware.

    Of course, there are also a few software providers that have invested heavily into their technology and have dedicated people and resources into ensuring a safe and secure election experience for all. The challenge for any government is to separate the wheat from the chaff: what software vendor can be trusted with something as important as an election, and which vendors are selling trumped up snake oil? At this point, even a lot of tech savvy folks can’t tell the difference.

    The solution to this conundrum is to establish some standards for internet voting. Rather than throw the baby out with the bathwater, let’s figure out exactly what IS needed to satisfy all the requirements for internet voting. How secure is secure? What kind of authentication is needed? What is the process for establishing auditability? I think we’ve finished the “study internet voting” phase, and we’ve made a lot of same conclusions. There’s also a lot of assuming going on, and that’s just frustrating those who are interested in embracing the technology. Does it improve voter turnout? Maybe, can’t say for sure. Does it save money? Possibly, but it depends. The only thing we can say for certain is that is does improve accessibility. The rest is influenced by the software and the services that the software vendor provides. So, OF COURSE there’s going to be a lot of variability in the results!

    Once we set some standards for the software providers to live by, they can go about their business of solving the workflows, do the testing, and submit their solutions for certification. If government can do that, then they will also create a much easier apples-to-apples comparison for pricing, service and quality – in addition to all the other deliverables for electoral trust. After that’s done, we can set some benchmarks for internet voting in elections, and we’ll also know what good looks like, and how to improve and optimize the experience for everyone.

  3. William J. Kelleher, Ph.D.

    Internet voting has been used in public elections all around the world over 100 times, w/o one security problem. At least 40 cities in Canada.
    Kinks in NDP and Washington DC were not real elections. For more info see http://tinyurl.com/IV4All “Internet Voting for All” blog.

  4. The preliminary report of BC’s independent panel on Internet voting has now been published. It too is sceptical of online voting and says it should not be permitted except for people with serious accessibility issues.

    Here is the main recommendation:

    Do not implement universal Internet voting for either local or provincial government elections at this time. However, if Internet voting is implemented, its availability should be limited to those with specific accessibility challenges. If Internet voting is implemented on a limited basis, jurisdictions need to recognize that the risks to the accuracy of the voting results remain substantial.

    It sets out its own criteria for evaluating online voting, which can be compared to those of Ontario’s Chief Electoral Officer set out in the main article above. The BC Panel gives these criteria (explained in more detail in the report):

    Accessibility
    Ballot anonymity
    Individual and independent verifiability
    Non-reliance on the trustworthiness of the voter’s device(s)
    One vote per voter
    Only count votes from eligible voters
    Process validation and transparency
    Service availability
    Voter authentication and authorization