Paper Vote Canada

After I tweeted a blog post response yesterday to an article about online voting and an associated radio interview, 1310 News contacted me for an interview.  I really appreciate their rapid outreach and the opportunity to speak.


I was on Ottawa Today with Mark Sutcliffe from about 11:50 to noon (Oct 26, 2016), the track is online now, to hear me go to timecode 34:33.  The interview ends at timecode 43:34.

The one thing I wish I had mentioned is that it's not just my personal opinion that every component needed for an online voting system is not secure (from your desktop or phone, through the network, to an elections vote counting server), it's the consensus evidence from computer scientists who have been researching this issue.  It is very unfortunate that in the upcoming McGill electoral reform event and in particular the upcoming uOttawa electoral reform event that has e-voting on the agenda, there are zero computer security experts presenting.  You wouldn't have an event about designing bridges without bridge engineers, why is it ok to have a discussion about using technology for voting without software engineers present?

In addition to this blog, you can find me on Twitter @papervote

Labels: comment, internet voting, links to audio, online voting

Here's my comment on the iPolitics article Canadians want online voting
(I have also left a similar comment on the EKOS Politics article The Case for Internet Voting,
and this can serve as a partial response to the claims made on 1310 News Ottawa Today with Marc Sutcliffe, Oct 25, 2016 from 10:00 to 11:00, timecode 11:10 to 21:02):

What I hope has roughly the same shelf life as whips and buggies is the idea that we can make tech policy based on opinion polls.  Before we send a spacecraft to Mars, we don't poll people for their opinions on whether the Earth orbits the Sun or on the laws of gravity.  In the 21st Century, we need to do evidence-based policy.  Here's what the computer science and social science evidence says about online voting:

• Widespread use of online voting would enable widespread coercion of voters, including vote buying.
• The innumerable software and hardware components that would be involved in marking, transmitting, receiving and counting an online ballot represent an unreasonably high risk to the chain-of-custody for the ballot.
• Canadian government departments have already been successfully cyberattacked by nation-states.
• Computer security experts warn that online voting is not secure.
• National security experts warn that online voting is not secure.
• Social science evidence indicates that online voting won't increase turnout.

Let's take turnout evidence first.  Just last week there was online voting again in Halifax.

Halifax Votes 2016: E-voting turnout down by more than 10,000 from 2012
http://www.metronews.ca/news/halifax/2016/10/13/halifax-votes-2016-e-voting-turnout-down-from-2012.html

In fact, the evidence shows that adding online voting never substantially increases turnout.  In a study of online voting in Ontario, Goodman and Stokes conclude

"internet voting is unlikely to solve the low turnout crisis"

Goodman, Nicole and Stokes, Leah C, Reducing the Cost of Voting: An Empirical Evaluation of Internet Voting's Effect on Local Elections (October 6, 2016). Available at SSRN: https://ssrn.com/abstract=2849167

Estonia, a tiny country that is often claimed as a poster child for online voting, had a lower turnout in its national elections in 2015 than Canada did.  (Estonia 64.2%, Canada 68.3%)

Claims that online voting will somehow engage young voters are similarly spurious; the evidence shows that it is middle-aged voters who would have voted anyway who vote online.  The City of Kitchener’s 2012 report on Internet voting finds that “There is clear evidence that, regardless of geography internet voting does not attract younger voters.” (FCS-12-191 - Alternate Voting - Internet Voting)

Similarly, on the security front, the article proceeds from one assumption unsupported by evidence to another.  First "If we can make financial transactions secure".  Well actually we can't make financial transactions secure.  The core interbanking system SWIFT was compromised and millions of dollars were stolen. 

New York Times - Once Again, Thieves Enter Swift Financial Network and Steal http://www.nytimes.com/2016/05/13/business/dealbook/swift-global-bank-network-attack.html - May 12, 2016

Online banking transactions are individually identifiable and reversible.  That's why online banking works.   When people's banking accounts are hacked - and they are hacked all the time - the charges are reversed by the bank at no cost to the customer.  Votes can't be either individually identifiable or reversible.  That's one of the core challenges of online voting.

Blockchain isn't a magic bullet for online voting.  You can watch Canadian computer scientist Jeremy Clark explain why https://youtu.be/-Uw-JnGIeIE

And it's important to understand hacking isn't some theoretical threat.  Core internet services were disrupted just this past week by a Distributed Denial of Service (DDoS) attack.

New York Times - Hackers Used New Weapons to Disrupt Major Websites Across U.S. - http://www.nytimes.com/2016/10/22/business/internet-problems-attack.html - Oct 21, 2016

Canadian government departments including the Finance Department, Treasury Board Secretariat, Defense Research and Development Canada and the National Research Council have all been successfully hacked. 

CBC - Foreign hackers attack Canadian government - http://www.cbc.ca/news/politics/foreign-hackers-attack-canadian-government-1.982618 - February 16, 2011

Globe and Mail -  Extensive fallout from Chinese hack of National Research Council - http://www.theglobeandmail.com/news/national/records-show-extensive-fallout-from-chinese-hack-of-national-research-council/article31695327/ - Sep 2, 2016

Online polls where you ask people whether they like doing things online are all well and good.  But they're not evidence supporting policy on online voting.  When the actual evidence is examined, the conclusion is that online voting is a terrible idea.

For more information for a general audience, see

* Why Electronic Voting is a bad idea https://youtu.be/w3_0x6oaDmI
* Online Voting is a cybersecurity nightmare http://www.dailydot.com/layer8/online-voting-cybersecurity-election-fraud-hacking/ - June 10, 2016
* This is why we still can't vote online http://motherboard.vice.com/read/this-is-why-we-still-cant-vote-online - Oct 19, 2016

For a technical audience, you can review the many very real vulnerabilities and successful attacks documented by computer security expert J. Alex Halderman in "Practical Attacks on Real-World E-Voting" https://jhalderm.com/pub/papers/ch7-evoting-attacks-2016.pdf

Labels: comment, internet voting, online voting

@kady has written an article in the Ottawa Citizen: Craving a break from the referendum-or-not debate? Why not fight about online voting instead?

I would have to turn on Facebook Platform apps to leave a comment on the Citizen, which isn't going to happen, so I'm putting my comment here instead:

COMMENT

Kudos to Chief Electoral Officer Mayrand for stressing caution.

I'd like to quote part of Mayrand's testimony that illustrates the many challenges with introducing online voting:
"...we have to be careful.  We need to look also at security.  There are certain characteristics of the vote that we want to preserve: it is confidentiality, it is secrecy, the reliability and the integrity."

There is extensive evidence that we do not currently have the technology to meet the security requirements needed for online voting, nor the ability to maintain the confidentiality we currently get with the paper secret ballot.

A 2014 report from the BC Independent Panel on Internet Voting ( http://www.internetvotingpanel.ca/ ) indicated that
"Internet voting has some significant inherent risks"
and recommended
"Do not implement universal Internet voting for either local government or provincial government elections at this time."

An extensive investigation conducted by the Parliament of Australia in 2013 ( http://www.aph.gov.au/Parliamentary_Business/Committees/Joint/Electoral_Matters/2013_General_Election/Second_Interim_Report/Chapter_4#safety ) indicated that
"Internet voting is considered by experts to be the most risky and difficult mode of electronic voting to implement."
and concluded that any potential benefits do not outweigh the risks.

You can find a lot more information about this topic in my blog "Paper Vote Canada" ( http://blog.papervotecanada.ca/ )

ENDCOMMENT

(The quote from Mayrand is from my manual transcription of his July 7, 2016 testimony, as the official transcript is not yet available.)

You can see previous comments I have left on newspaper articles by selecting the label: newspaper comment.

Labels: comment, internet voting, newspaper comment, online voting

Here is the comment I posted on the Financial Times article A vote for online elections (June 10, 2016)

COMMENT

Considering that the Financial Times itself reports "The tools for cyber attacks are so accessible that individuals and private groups, as well as states, can carry out such offensives" it seems extraordinarily unwise to open up voting to these kinds of attacks.  (FT quote from Germany points finger at Kremlin for cyber attack on the Bundestag http://www.ft.com/intl/cms/s/0/668a131e-1928-11e6-b197-a4af20d5575e.html )

Sophisticated and in some cases state-sponsored attackers have broken in to White House computers and the Canadian National Research Council.  We know this because some of the breaches have been detected and (to the extent possible) repaired.  (BBC News - White House computer network 'hacked' http://www.bbc.com/news/technology-29817644 ; BBC News - Canada National Research Council 'hacked by Chinese spies' http://www.bbc.com/news/world-us-canada-28548925 )

Elections cannot be repaired after the fact, and they shouldn't have to be. Online voting increases risk enormously, without providing any benefits.  Even if online voting security were extremely good, there would still be the issue of coercion.  But time and again when voting technology is examined, the security is inept at best.

The Estonian example is illustrative, for while "officials say that their decade-long experiments have gone very well", an independent report on e-voting in Estonia identified serious concerns including "staggering gaps in procedural and operational security, and the architecture of the system leaves it open to cyberattacks from foreign powers, such as Russia" (Independent Report on E-Voting in Estonia, https://estoniaevoting.org/ )

It is key to choose appropriate technology, not just technology for technology's sake.  The paper secret ballot is still the appropriate technology for voting, as no computer system has yet been created that can provide the same combination of security and privacy.  I realise this may "seem peculiar" but it is the reality, as leading computer security experts have repeatedly stated, including most recently David Dill of Stanford University (Stanford Engineering, Why Online Voting is a Danger to Democracy http://engineering.stanford.edu/news/david-dill-why-online-voting-danger-democracy )

ENDCOMMENT

permalink to comment is http://on.ft.com/25RBV2F - it was posted June 11, 2016

Labels: comment, internet voting, newspaper comment, online voting

Here is the comment I posted on The Economist article Online voting could transform Britain’s electorate - April 18, 2015

COMMENT

Labels: comment, internet voting, newspaper comment, online voting, UK

Here is the comment I posted on The Australian article Election cliffhanger prompts e-Voting rethink by Supratim Adhikari on July 5, 2016.  The comment has not (yet?) been accepted.

COMMENT

While the counting of paper ballots is slow, all of the complexity is visible to everyone.  With an Internet voting system, you would be greatly increasing the complexity and all of the associated possibilities for very complex systems to fail, but hiding all of that increased complexity from voters.  This simply distances the voting system even further from citizens.  To quote the 2013 Parliamentary inquiry: “It is important that in embracing technology, the secret ballot is not undermined, voter behaviour is not negatively impacted, and confidence in the electoral process and electoral outcomes is not damaged. At a time of debate about community disengagement with political processes, it would be greatly concerning if the method of voting—the one act of participatory democracy that all Australian citizens will definitely engage in—was to further disengage the community from these processes.” [1]

That is not to mention the risk of the election being hacked, including by other nation states.  This is a reality already - China was blamed for a massive hack on a Bureau of Meterology computer [2] resulting in the government spending an additional $230 million on cybersecurity [3] .

This leads to the last point, about claimed cost savings.  The current paper-based system can be used again and again, year after year, with at most minor modifications.  That is not at all the case for any kind of universal electronic voting system.  Technology and cyberattacks are constantly changing, which means that any electronic voting system is obsolete the day that it launches.  Such a system would have to be continuously and expensively updated for every election, to reflect technology changes and to defend against new types of attacks.  That kind of continuous expenditure on information technology is great for the Information Industry, but terrible for government budgets.

[1] http://www.aph.gov.au/Parliamentary_Business/Committees/Joint/Electoral_Matters/2013_General_Election/Second_Interim_Report/Chapter_4#comment section 4.80
[2] http://www.abc.net.au/news/2015-12-02/china-blamed-for-cyber-attack-on-bureau-of-meteorology/6993278
[3] http://phys.org/news/2016-04-australia-hack-boosts-cyber.html

ENDCOMMENT

With the slow paper-based count, there has been a flood of articles about Australian e-voting.  Unfortunately I'm not able to respond to them all, but the above comment applies generally.

Labels: australia, comment, electronic voting, internet voting, newspaper comment, online voting