4 challenges of remote electronic voting

Remote electronic voting

Every election brings renewed discussion of remote electronic voting, or online voting: allowing people to vote from home without travelling to a polling station and queuing. It could also save environmental resources and perhaps money: polling stations would cost less, but the system would need maintaining. I decided to briefly explain why I do not think universal online voting is currently realistic.

Technical challenges

The technical challenges are, objectively, the easiest to address. They mainly come down to resilience and reliability.

Resilience

An online voting system must guarantee availability throughout the election, avoiding slowdowns and outages. Near-100% service availability, potentially 99.9%, requires redundancy in every component and scalability to prevent bottlenecks.

Reliability

The service must also be completely reliable. Errors in collecting and counting votes are unacceptable. This is less trivial than it sounds, but with appropriate investment and resources, it is reasonable to expect 100% of votes to be correctly collected and none lost, even remotely.

Rights-related challenges

Protecting the rights required by Italy’s electoral system presents greater obstacles. Some cannot be guaranteed with current technology without greatly complicating the process. Article 48 of the Constitution defines voting as personal, equal, free and secret, and its exercise as a civic duty.

Personal

Voting cannot be delegated, except for certified physical impediments. Every voter must therefore be able to use the necessary tools independently. Italy has low digital literacy and a wide digital divide.

Some studies, such as this report, identify areas where up to 30% of citizens do not use or cannot access the internet. Bridging this requires continued infrastructure investment and, especially, investment in digital literacy. Access to public services can be delegated to relatives, friends or assistance centres; constitutionally, voting cannot.

Equal

Every person has one vote. Digital identities can ensure this without difficulty.

Free

Simply logging in with SPID or CIE, Italy’s digital identity systems, does not readily guarantee a freely cast vote. People could vote under threat, blackmail or payment. Proving that a purchased vote was cast as requested would be far too easy.

How could we guarantee that a relative or criminal is not at home forcing someone to vote a particular way?

Secret

Running the system would require storing substantial data. Who could guarantee that analysis could never link a digital identity to a vote? Guaranteeing complete secrecy alongside voting security would be one of the toughest challenges, though technology might solve it.

The human challenge

People resist change. We recently saw this with the Immuni app: protecting all the required rights meant restricting its tracking features and potential until it was not fully effective. Resistance also creates a need to earn citizens’ trust that their constitutional voting rights will be respected. Consider the annual arguments over using a pencil to mark ballot papers, or people legitimately asking to put their own ballot in the box. Many edge cases are manageable in conventional elections but difficult remotely.

Digital literacy, already mentioned, is another major issue.

According to a recent ISTAT survey, almost 340,000 people in Italy were illiterate. Asking someone to mark a party symbol is possible; requiring them to obtain SPID, operate a digital device, navigate a website and complete authentication and voting procedures is much harder.

Security

This could be classed as a technical challenge, but deserves separate attention. Security goes beyond computers. As discussed, remote voting could open opportunities for criminals to gain power and money, potentially endangering vulnerable citizens who are easier to manipulate.

Any internet-facing system can suffer cyberattacks and data breaches. A high security standard is achievable, but finding someone willing to guarantee 100% security across the entire chain would be difficult.

Alternatives

One reason people demand online voting is wasted resources. Setting up polling stations and printing large quantities of material has significant financial and environmental costs. A first step could simply be digitising the vote itself: electronic voting equipment, as used in other countries, including some US states, while retaining physical attendance and most existing procedures. Protecting voters’ rights would be much easier. The equipment would not face the internet, and tampering would require physical access.

Trials so far

Estonia

Remote electronic voting was first used in 2007, with just over 5% of eligible voters choosing the internet. European observers were satisfied with the experiment, but highlighted some of the concerns discussed here. See this account.

By 2014, almost 30% of eligible voters voted online.

Switzerland

Trials were halted in 2019 over voting security weaknesses. They resumed in 2023, in some cantons only for residents abroad, and in Basel-Stadt also for people with disabilities and selected pilot municipalities. Access codes were sent for the voting platform.

Norway

Trials began in 2003, but after extensive parliamentary debate over security, all online voting was suspended from June 2014.

Germany

Electronic voting tests ran between 2000 and 2006. In 2009 the Constitutional Court blocked them as incompatible with proper electoral procedures.

France

Residents abroad began being able to vote remotely in 2003.