In a fully functioning democracy, the Covid pandemic has posed a formidable challenge to people exercising their collective rights as citizens, namely the right to vote.
Large public gatherings associated with elections raise serious public health concerns that could lead to the virus spreading rapidly. Such was the case with the Sabah state election in September last year, when we witnessed an exponential spike from 82 cases on 26 September (polling day) to 1,240 cases just a month later. It proves that an election can be a “superspreader” event for the virus.
With the Malacca state election on 20 November, many are undstandably that it could be another superspreader event and spark another outbreak. Hence the rising public clamour to change the way we normally conduct elections with the primary goal of avoiding large gatherings of people, especially in confined spaces. One suggestion is to introduce electronic voting or e-voting.
E-voting allows voters to cast their ballots over the internet from their electronic devices, typically smartphones and computers. Presently, only a few places in the world have implemented e-voting or experimented with e-voting for elections of varying levels.
Estonia, a Baltic nation with fewer people than Penang, is the only place that has carried out e-voting for its national election since 2005. Other countries, such as the US and Switzerland, limit e-voting to specific localities and voting groups (overseas military voters, for the US).
Many places have abandoned e-voting after pilot runs during local elections. Most countries do not even consider e-voting as an option for high stakes elections.
Why conventional voting is more secure
So why is e-voting is not commonly implemented across the world, especially in technologically advanced countries that also happen to be fully functioning consolidated democracies?
Why hasn’t e-voting taken off in popularity and ubiquity like other electronic-based transactions that are now part and parcel of our lives, such as e-commerce, e-banking and e-government?
Two key aspects of a trustworthy election are secrecy and verifiability, which are mutually dependent.
The ballots that we cast as voters are secret, meaning only the voters themselves know the choices made on those ballots and those choices would remain unaltered by the time they reach the counting process.
Unlike other types of electronic transactions such as e-banking and e-commerce, which offer full verification on both ends of the transaction, e-voting only verifies voters before they cast their e-ballots.
Once e-ballots enter the system, they remain anonymous. Election officials at the end of the e-voting transaction cannot verify if those e-ballots are cast by the same voters or if the choices made on the e-ballots represent the true intention of the voters. Besides voters’ verification, the voters themselves would want to verify that their ballots – and, by extension, the choices they have made – are accepted by the election officials.
In a conventional voting procedure, election officials would ask voters for identity documents to verify their identity before handing them the ballots. After making their choices, voters would submit their ballots to the election officials and the ballots are then stuffed inside a box or run through an optical scanner. In short, verified voters get to observe with their own eyes and through their own physical actions that the integrity of their ballots remains uncompromised throughout the voting process.
At present, e-voting cannot offer a similar degree of confidence in the secrecy and verifiability of the vote compared to conventional voting. Our public internet infrastructure is still highly vulnerable to outside interference. In e-voting, interferences can take place at three points of attack, from the moment voters cast their e-ballots until election officials count those e-ballots.
Vulnerabilities of internet infrastructure
The first point of attack, also the weakest point in the e-voting ecosystem, is the voters’ own electronic devices (the front end). Unbeknown to voters, malware – software designed to wreak havoc and steal private information – might have infected their electronic devices.
In 2020, there were 160 million malicious programs (malware) and 83% of these malware targeted Windows-based PCs. Malware in voters’ electronic devices can manipulate choices voters make on their e-ballots or attach themselves to e-ballots to infiltrate the central database used by election officials to store the votes.
To put in a real-world scenario, because of severe malware corruption, the choices that voters make would not match the choices ultimately counted by election officials. Imagine millions of compromised e-ballots that do not represent the true intention of voters and consequently its impact on the legitimacy of the election and the democratic system itself.
The second point-of-attack can occur when e-ballots shuttle between the front end (users’ electronic devices) and the back end (the central server used by election officials). In what computer security experts call “man-in-the-middle” (MitM) attacks, hackers can attach malware onto e-ballots mid-travel or reroute e-ballots through legitimate-looking phishing websites to harvest private confidential information.
In either case, voters’ e-ballots are severely compromised, putting the results of the election in serious doubt.
The third point-of-attack is the central server used by election officials to store all the e-ballots cast by voters. The likeliest type of attack deployed by hackers on the central server would be a “distributed denial of service” (DDoS).
DDoS attacks occur when hackers overload the server with “requests”, slowing down and ultimately crashing the server. This can be carried out by deploying armies of bots to flood and overwhelm the server. In e-voting, DDoS attacks can severely disrupt the efficacy of the voting process by causing repeated failures when voters try to log in their votes into the central server. This frustrates the voters and calls into serious question the legitimacy of the election results.
Magnitude of voting irregularities
What separates online and offline voting irregularities is the scale of manipulation. In offline voting, postal ballots remain the most vulnerable to tampering because of the indirect way that these ballots travel from voters to election officials.
Still, incidents of manipulation of postal votes are localised and individualised because of the logistical dispersion of the mailed-in votes. Simply put, it is next to impossible to orchestrate postal ballot tampering on a massive scale that involves tens of millions of votes.
The same cannot be said for e-voting. The aggregation of e-ballots in central servers means vote could be tampered with on an unthinkable scale, to the point where the election results might even be nullified.
It is also difficult to convince voters that the results of a closely contested election are fair and legitimate because of the sheer complexity of technology that makes up e-voting and the inability to verify the original choices made in the ballots.
At the heart of it are the paramount questions of trust and reliability – whether the e-voting system can fully guarantee ballot secrecy and integrity of voters’ intention. These are the questions that e-voting has yet to address satisfactorily.
The ideal e-voting system
The currently unattainable standard of secure and trustworthy e-voting is called “end-to-end verifiability in i-voting”, better known by its acronym E2E-VIV.
The objectives of E2E-VIV are threefold: verify that voters’ votes are included in the election outcome; verify that the system records the content of the votes correctly; and verify that the number of people who vote for a given candidate is accurately counted.
The three objectives can be summed up in a simple phrase: “cast as intended; recorded as cast; and counted as recorded.”
If we are to implement e-voting in light of our current technical shortcomings, what efforts can be taken to achieve the objective?
Of course, we need to ensure that the public internet infrastructure is safe from nefarious outside interference, which at this moment still contains plenty of exploitable loopholes, as stated above.
One way to circumvent this structural weakness is to disconnect the e-voting ecosystem from the public internet infrastructure. This can be achieved by only carrying out e-voting at polling stations or designated kiosks in selected government buildings. However, it means that voters would still have to make their way to public places to cast their votes, which is not ideal during a pandemic.
Sadly, when it comes to e-voting, this is as close as we can get to it, given the set of technical challenges we are facing now.
How to make voting more accessible now
E-voting is not the only available option to improve voting accessibility. There are other ways to expand enfranchisement among the voting-age population, whether they are in the country or overseas.
One way is to provide eligible voters with the choice of voting by mail, along with the conventional option of voting at the polling stations on election day. This practice is known as “no-excuse absentee voting”.
This is mainly done to facilitate convenience, allow cost-saving measures, ease the pressure on polling station staff on election day, and reduce congestion and wait-times at polling stations on election day. Postal voting need not be limited to Malaysian voters living abroad but can be used by all Malaysians without exception.
Many countries also allow voters to cast their ballots ahead of election day, known as in-person advanced voting. In contrast to postal voting, in-person advanced voting needs to be carried out at specially designated polling sites such as government buildings, schools, places of worship, hospitals and other public places.
In a typical in-person advanced voting process, election officials reserve a period before the election day for advanced voting. The period can range from a few days to a few weeks. Voters can visit any of these specially designated polling sites during the early voting period to cast their ballots.
The ability to vote early provides voters with convenience and flexibility while reducing congestion at polling stations and pressure on poll workers on election day.
Other useful suggestions gleaned from a recent BFM Radio Twitter poll include having separate voting days for MyKads ending in even and odd numbers to reduce congestion at polling stations.
The Election Commission can also set up “remote voting” polling stations in every state for out-of-state voters, so they don’t have to travel back to their home towns to cast their ballots.
Many studies have shown that voter turnouts would increase significantly if the voting system is made more accessible and hassle-free. A democratic government should always aim to maximise voter turnout as it also strengthens its own legitimacy to govern.
The key point is that there are other ways to adapt our voting system to these extraordinary times – ways that have been tried and proven in many countries around the world.
The government has commendably enfranchised a sizeable group of voters by passing the Undi 18 law (to reduce the minimum voting age to 18). It can now go a step further by reforming the voting system to make it more accessible, safe and convenient for the public, as suggested above.
The day when we can vote while watching Netflix on the sofa has not yet arrived, but rest assured that time will come soon enough.