Introduction and Definition
The COVID-19 virus pandemic has raised global concerns about the security of traditional face-to-face voting. This has led to renewed interest in remote voting technologies, in particular e-voting.
E-voting is an umbrella term that covers many types of voting. E-voting is defined as any form of voting that uses electronic or computerised means to speed up or simplify the process of expressing a preference. Electronic voting techniques, also known as e-voting or electronic voting, consist of tools that allow votes and preferences to be expressed using electronic technologies. These include punch cards, optical scanning systems and direct optical recording systems. The last option is the most developed in the world and allows the voter's preferences to be stored directly by recording the data on storage media using touchscreen tools (Russel & Zamfir, 2018).A computer is placed in each voting booth and the voter can express their preferences through a touchscreen or by using buttons on the sides of the monitor. The monitor is often an exact replica of the paper ballot to make voting easier. The computer prints out a ballot with the selected preferences so that the voter can check them before the printed ballot (or transaction receipt) is automatically deposited by the system in an inaccessible container.
Electronic voting, like traditional voting, is divided into manned and unmanned voting. Supervised voting takes place in public and supervised polling stations, while unattended voting can be carried out remotely, at any place, from any terminal and without any form of supervision at the time of voting. The former can be equated with voting on touch-screen devices in the traditional polling station and does not require a network connection (offline voting) or may use a dedicated network, while the latter can consist of voting via the Internet (or even by SMS) from any electronic terminal connected to the network (online voting).
Electronic voting has been subject to scrutiny, particularly in relation to political and electoral consultations. Ron Rivest's statement that "voting is a hard problem" (Belanger, 2020) provides a succinct picture of the difficulties of managing any electronic voting system. First and foremost, it is necessary to guarantee security procedures that allow the voter to cast his vote in full respect of his privacy.
More generally, voting procedures must necessarily address the following ethical, legal, and societal challenges (National Academies of Sciences, Engineering, and Medicine, 2018)
- Protect the secrecy of the ballot,
- ensuring the recognition of voters to allow them access to the system, while at the same time ensuring that votes cannot be linked to voters, i.e. protecting the anonymity of each voter's preference,
- ensuring the possibility of remote voting for voters who are outside their area of residence on the day of the vote,
- ensuring that the voting system is clear and understandable, so that it can be used by all voters (including those who are illiterate or disabled),
- to ensure that all voting and counting operations are carried out automatically, without the use of paper, in order to prevent the falsification or invalidation of votes or any part of the electoral documentation,
- and, last but not least, to reduce the time and cost of the electoral system.
The identification of the voter, who is provided with a polling card and an identity document, is normally carried out at the polling station through the physical presentation of the voter, thus eliminating the possibility of false identity and multiple voting. In the case of e-voting, recognition could be facilitated by electronic ID documents, which would, among other things, avoid transcription errors in personal data and allow voting at any polling station. Some scholars (Russel & Zamfir, 2018) have also argued that e-voting could encourage participation in elections, thus bringing citizens closer to exercising one of the most important rights provided by democracy.
The issue of vote counting concerns the stage where the greatest time savings would be recorded. Each ballot would be counted by the electronic processors in a few seconds and immediately transmitted to the control bodies. In this way, the results would be available a few minutes after the polling stations closed.
Another form of e-voting is optical scan voting, where the voter marks a ballot paper and inserts it into a machine. The votes are checked by automatic sensors at a central location or at the polling station itself. If the vote is scanned at the polling station itself, the system also checks that the vote is valid at the moment it is accepted. Any problems with optical scanning systems could be caused by extraneous marks on the ballot paper, marks inadvertently placed outside the spaces provided, marks made with incompatible writing equipment or incomplete marks.
Difference between e-voting and i-voting
Another method of e-voting is called i-voting, also known as Internet voting. Unlike e-voting (electronic voting or manned voting), which refers to a computer installed in a polling station that speeds up the voting process but still requires the physical presence of the voter at the polling station, i-voting (Internet voting or unmanned voting) involves the possibility of voting remotely via an Internet connection. In other words, i-voting is a voting method that uses the Internet and, through the use of remote stations, allows voting from any location equipped with a device with an Internet connection and a web browser. Voters can be identified by authentication on the website or by other methods (e.g. national identity card, smart card, token, etc.). The authentication process is carried out using a PIN and a TAN, which are provided with security keys and asymmetric encryption, designed to ensure that the expression of the vote is secure and secret, while respecting the fundamental principles of voting. Once the voter arrives at the site, he or she must prove his or her identity through the above-mentioned systems, after which the web system presents the electronic ballot on the screen; the voter expresses his or her preferences and verifies their correspondence on the screen. The voter's confirmation transmits the completed ballot to the election server.
Authentication is as important in Internet voting as it is in paper ballot voting at the polling station. The voting system can be connected to a network of polling stations with web, which use electronic ballot papers and transmit the voting data from the polling station to a collection point. It is possible to transmit the data at the moment each ballot is cast, or periodically, in groups of data, throughout the election period, or again in a specific group, at the end of voting. The results are therefore either sent out already counted centrally or, in the case of centralised voting, at a single collection point.
Internet voting systems are currently used in many countries, both in the private and public sectors. In the United States of America, the United Kingdom, Estonia and Catalonia, e-voting has also been used for general elections and referendums. Switzerland has also been using i-voting for local referenda for a number of years. Voters receive a password to access the ballot paper by post. For online e-voting in local elections, there are various platforms where all eligible voters can participate without having to go to the polling station in person (Darmawan, 2021). The benefits of i-voting include reducing costs, speeding up processes (such as vote counting) and reducing organisational time. The PIN and TAN authentication process with security keys and asymmetric encryption also ensures that the vote can be cast securely and secretly, respecting the fundamental principles of voting.
Computer security and the protection of personal data are paramount in i-voting methods. It is necessary to guarantee the authenticity of the preferences expressed by voters in order to ensure an accurate final count, and to ensure that each vote is recorded without the possibility of being changed. It is also essential to eliminate the possibility of voters exploiting imperfections in the system to vote more than once, thus ensuring robustness and reliability to guarantee the exact correspondence between the votes cast and the votes recorded (Schryen & Rich, 2009).
Finally, it is essential that the voter cannot be associated with the vote cast (so-called voter anonymity), especially during the counting phase, which must be organised in such a way as to make it impossible to link the vote to the voter. In this sense, the specific statement of the Committee of Ministers of the Member States of the Council of Europe in its Recommendation CM/Rec (2017)5 (Council of Europe - Committee of Ministers, 2017) on e-voting, according to which "votes are and remain anonymous", is instructive.
E-voting processes in the world and in Europe
The concept of e-voting does not yet have a clear international definition. As already mentioned, the term e-voting describes a number of different approaches to automated techniques applied to the electoral process, i.e. procedures that are quantitatively and qualitatively very different from each other (Rapporteur Group - CM - COE, 2017). Generally speaking, e-voting refers to the process and result of applying new computer technologies to the electoral process. However, these may have different degrees of application and may also be included in different phases of the electoral process; in fact, the use of information technologies may concern, as a whole or individually, the preparation of the electoral roll, the identification of the voter, the expression of the vote, the ballot and the transmission of data (ComputerScience.org Staff, 2022).
The first form of e-voting took place in America with the punch card, and currently in the United States, even within a single state, there can be as many as five different types of voting for presidential elections: the paper ballot, the punch card, the optical reader, the punch and the actual electronic vote. Optical reader systems and direct recording electronic systems can either provide a paper receipt to prove the vote or be fully digital and able to identify the voter, who has a smart card, through a POS-like interface. However, all these forms of e-voting require the presence of the voter, who must be physically present at the ballot box and recognised by the electronic system before casting a vote (ComputerScience.org Staff, 2022). More recently, Internet voting has gained popularity in general and local elections in Canada and in primary elections in the USA. In Brazil, since 2000, citizens have voted using computers located at polling stations, and their votes are stored locally in electronic ballot boxes until the votes are counted. Electronic voting machines are also being used extensively in India (Darmawan, 2021).
In Europe, many countries have experimented with e-voting, with mixed results, including the UK, Ireland, the Netherlands, Norway, Germany, Switzerland, Spain, Estonia and Finland. Online voting continues to face obstacles and barriers, mainly due to unresolved security and ethical issues. After years of experimentation, Norway finally says no to 'e-voting', calling it 'a useful experiment, but only traditional polling stations can guarantee the freedom and secrecy of the vote', basically for two reasons. The first is related to the 'security' of e-voting, in fact many Norwegian citizens expressed fears and concerns about e-voting. In particular, the greatest fear was that their political preferences would be published on the Internet, especially on smartphones and tablets, in the event of hacking or something similar. The second reason, however, is the failure to increase voter turnout, including among younger voters. In 2013, for example, out of 250,000 citizens invited to vote electronically, only 70,000 expressed their preferences online (Gjøsteen & Lund, 2016).
From the late 1990s until 2007, electronic voting was the norm in the Netherlands, but it turned out that the direct optical recording systems could be manipulated relatively easily without the election offices noticing, which led to a major scandal and an immediate decision to return to the good old ballot paper. In the case of the 2006 Dutch parliamentary elections, both main types of electronic voting were used: via the Internet for those abroad and via the direct optical scan systems in some polling stations. The choice between optical scan machines and paper ballots was left to each municipality, and even within the same municipality different polling stations could have a different voting method. The vulnerability of the software raised real doubts about the possibility that the election could be manipulated, and the Dutch Ministry of the Interior was forced to respond when faced with a hacker attack on the national computer system (Gibson, Krimmer, & Teague, 2016).
The organisation and administration of federal elections in Switzerland varies greatly from canton to canton and municipality to municipality, and e-voting is currently being tested. After more than 200 successful trials, a total of fourteen cantons have allowed part of their electorate to vote electronically since 2004. Today, ten cantons offer e-voting. In five of them (FR, BS, SG, NE, GE), both Swiss abroad and Swiss residents can vote; in the remaining five (BE, LU, AG, TG, VD), only voters resident abroad can vote electronically. (Gibson, Krimmer, & Teague, 2016).
In Germany, e-voting was introduced in 1999 and was not widely used until 2005. However, the German Supreme Court declared the use of e-voting, a method used for two million voters in the 2005 elections, to be 'unconstitutional', thus returning to ballot papers and pencils. The German court's decision was based on the 'unreliable nature' of the machines used and the 'black box' system that makes the voting process non-transparent and open to manipulation, especially at the counting stage: as the counting process is secret, no one can be sure that their electronic vote has been calculated correctly (Seedorf, 2015).
In Estonia, most voters are free to choose to vote via the internet in both local and parliamentary elections, as many of the eligible voters have access to the internet voting system thanks to the introduction of an ID card with a computer-readable microchip, and with these credentials they can access the ballot paper online. All Estonian voters need is a computer, a PIN code and an electronic ID card reader to vote from anywhere with Internet access and a web browser.
A special feature of e-voting in Estonia is that it can only be used on early voting days. This rule helps to ensure that the right to vote is exercised freely. In the event that the voter has inadvertently or under coercion made his or her preference clear during remote voting, the voter can go to the physical polling station in a second stage to exercise the traditional right to vote using a paper ballot, which effectively cancels the preference expressed through early e-voting (Vassil, 2016).
Problems related to the use of e-voting also occurred in France, where a project to introduce manned e-voting in 2003 experienced several inconveniences: the number of voters allowed to vote did not correspond to the number of votes cast, and the waiting time to vote was longer than expected. These drawbacks, among others, led the government in 2008 to reduce the number of municipalities allowed to use e-voting, which is still only used for expatriates (Darmawan, 2021).
In 2008, e-voting (generally manned) was also abandoned in the UK after numerous trials revealed a lack of public confidence in its accessibility and technical difficulties (Gibson, Krimmer, & Teague, 2016). In contrast, in three municipalities in Finland, the Supreme Administrative Court was forced to re-run the election in 2009, this time using paper ballots, after errors were found in the electronic ballot. However, both the UK and Finnish governments decided to rerun the internet ballot after deciding not to use electronic technology for voting (Darmawan, 2021).
Full digital citizenship presupposes effective computer literacy, i.e. a skilful use of technologies, as well as the possession of material resources and a technical and human capital capable of accessing and actively participating in political life through digital platforms. For these reasons, it is necessary to guarantee equal pre-conditions of accessibility to the Internet and to enjoy complete, impartial and neutral information.
Access to the Internet, developing a daily, virtual and online participation, besides allowing a fast flow of information, constitutes a new legal paradigm of the right to inform, to be informed and to be informed, as well as a modern opportunity to express one's ideas not only through the traditional - constitutionally guaranteed - means of expressing one's opinions. Information and accessibility to political, social, economic and cultural pluralism, as enhanced by today's techniques of mass communication, provide an opportunity to nourish a given order with democracy, measuring its level of openness and involvement and contributing to placing the citizen as such at the centre of public dynamics.
It is obvious at this point, however, to note a considerable disparity between those who have more or less access to e-democracy platforms: for example, between young and old, rich and poor countries, between educated and uneducated, between political activists and more. Digital inequality based on all these components leads to the fragmentation of societies and to the inevitable creation of a gap that leads to social and political hierarchisation of an electronic nature, resulting in undoubted discrimination and new and modern forms of inequality.
A low and inadequate level of participation causes the general interest to be neglected in favour of the particular one, since decisions would be taken only by that part of the population capable of using information technology. In addition to this, further problems arise when accessing IT platforms and network control over the data and information that can and must circulate becomes necessary.
There is the possibility, unfortunately, on the part of digital controllers, to disseminate only a particular type of information, not necessarily truthful, thereby conditioning the virtual citizen. A lack of computer literacy and the difficulty of verifying the truthfulness, correctness, completeness and neutrality of information leads to its manipulation, constructing a distortion of reality that becomes superficial and self-referential to such an extent as to preclude the formation of a mass digitised community that is effectively aware and critical.
These complications become even more evident when digital citizens are called upon to express their preference on political issues. They would, even if explicitly consulted, provide personal indications that could hardly adequately consider the criteria for the overall optimisation of the common good as a result of particular technical choices. Leaving this phase of decision-making directly in the hands of the citizens would constitute a risk, which would be further aggravated by the possibility of de facto ousting the 'digital illiterates', i.e. those people who have no computer skills whatsoever. For this hypothesis, the risk could be that of a predetermination of the questions and answers congenial to those who formulated them, with a consequent restriction of the faculty of choice limited to a 'yes' or 'no'.
The approach to support the activity of the policy-maker, i.e. to indicate operational choices, then requires extremely careful and sophisticated management of the results, capable of allowing a reading unaffected by statistical distortions. Moreover, if this tool were to be used for topics requiring highly specialised and evolved technical expertise, it could provide results without real meaning.
Political participation, in this way, becomes artificial since it resolves itself into a mere ratification of pre-packaged options, without any minimum opportunity for discussion and confrontation. Supporters of e-democracy believe that electronic procedures would speed up the entire electoral process by favouring a rapid and real-time counting of votes by exploiting the most modern and efficient technologies. They would thus increase voter participation by giving more voters the opportunity to cast their votes because of the assumed ease of using computer equipment.
Electronic voting would be expressed either from home computers or in manned public polling stations where the secrecy and personality of the suffrage would in any case be ensured by the identification of the voter by means of a pin, password or smart card, thus further expressing the democratic character of a state and a new mode of political participation. It must be borne in mind, however, that the importance of the right to vote, both in terms of its value and its historical-legal significance, cannot be totally entrusted to electronic devices that are liable to pollute the electoral competition without having the certainty of ensuring the most scrupulous respect for constitutional guarantees. The critical nature of this type of voting, in fact, is to be found in the resulting electronic equipment subject to the possible alteration of the electoral competition and, consequently, the lack of absolute protection of respect for constitutional guarantees.
Starting from this assumption, therefore, it can be deduced that home voting also presents serious criticalities inasmuch as not only would it not allow full identification of the voter in the freedom to vote, but it would also fail to ensure secrecy. There would also be a lack of certainty as to the concrete transparency of suffrage, its registration and its actual manifestation, since, as already mentioned above, one can never be certain of the proper functioning of hardware and software, just as one cannot exclude the possibility of hacking and manipulating the vote. The possible exclusion of those who do not have a high level of computer and technological education would also weigh heavily, both because they belong to distant generations and because of the difficulty of understanding the world of the net. Home voting, among other things, is, according to many, to be ruled out because of the possible use of devices that are used daily for other purposes, and this would conflict with the public function that only the presence at the polling station of the individual renders fully the character of a public subject".
An intrinsic weakness of such a system, in fact, is precisely the use of the voters' personal computers. The voter's terminal (client) is in fact one of the biggest problems in terms of security, as it is often inadequately protected against the most common cyber threats. Consider that in 2014, it was estimated that around a third of computers worldwide would be infected with malware, giving unauthorised parties the ability to spy on the voter, or even vote for them. Despite the potential benefits offered by online voting, therefore, including the possibility of increasing voter turnout, decreasing costs and improving accessibility of voting, progress in its implementation has so far been rather slow due to its vulnerabilities which also insist on the client/server system, the underlying structure of an internet voting solution.
In addition, it is now generally agreed that 'the human factor is always the weakest link in cybersecurity'. Voters need to be aware of the security threats to their device, hence, access to a secure and reliable device is the only credible option for certain results from elections held with i-voting systems.
Taking into consideration the server through which the voting service is offered, it is useful, in fact, to remember that an i-voting system can be subject to cyber attacks that can employ the most disparate methods.
Belanger, A. (2020). 3 Questions: Ron Rivest on trusting electronic voting systems. Retrieved 2022 from MIT News: https://news.mit.edu/2020/3-questions-ron-rivest-trusting-electronic-voting-systems-0226
ComputerScience.org Staff. (2022). Online Voting: Security Threat or Wave of the Future? Tratto da ComputerScience.org : https://www.computerscience.org/online-voting/
Council of Europe - Committee of Ministers. (2017, June 14). Recommendation CM/Rec(2017)5. Tratto da https://rm.coe.int/0900001680726f6f
Darmawan, I. (2021). E-voting adoption in many countries: A literature review. Asian Journal of Comparative Politics, 6(4), 482-504. doi:doi:10.1177/20578911211040584
Gibson, J., Krimmer, R., & Teague, V. (2016). A review of E-voting: the past, present and future. Ann. Telecommun., 71, 279–286. doi:https://doi.org/10.1007/s12243-016-0525-8
Gjøsteen, K., & Lund, A. (2016). An experiment on the security of the Norwegian electronic voting protocol. Ann. Telecommun., 71, 299–307. doi:https://doi.org/10.1007/s12243-016-0509-8
National Academies of Sciences, Engineering, and Medicine. (2018). Securing the Vote: Protecting American Democracy. Washington, DC: The National Academies Press. doi:https://doi.org/10.17226/25120
Rapporteur Group - CM - COE. (2017, April 20). Guidelines on the implementation of the provisions of Recommendation CM/Rec(2017)5 on standards for e-voting. GR-DEM(2017)10. Tratto da https://search.coe.int/cm/Pages/result_details.aspx?ObjectId=0900001680703319
Russel, M., & Zamfir, I. (2018). Digital technology in elections: efficiency versus credibility? - EPRS Briefing. Retrieved 2022 from European Parliament - Think Tank: https://www.europarl.europa.eu/thinktank/en/document/EPRS_BRI(2018)625178
Schryen, G., & Rich, E. (2009). Security in Large-Scale Internet Elections: A Retrospective Analysis of Elections in Estonia, The Netherlands, and Switzerland. IEEE Transactions on Information Forensics and Security, 729-744. doi:doi: 10.1109/TIFS.2009.2033230
Seedorf, S. (2015). Germany. The Public Nature of Elections and its Consequences for E-Voting . In A. Maurer, & J. Barrat, E-Voting Case Law: A Comparative Analysis. London: Routledge. doi:https://doi.org/10.4324/9781315581385
Vassil, K. (2016). The diffusion of internet voting. Usage patterns of internet voting in Estonia between 2005 and 2015. Government Information Quarterly, 33(3), 453–459. doi:doi: 10.1016/j.giq.2016.06.007