A revival of technology-assisted voting in NSW is likely to require more comprehensive legislative changes than first thought.
The final review into technology-assisted voting (TAV) was published with little fanfare about a fortnight ago [pdf], and contains some significant changes compared to an interim version published in August.
The major change is an expansion of the amount of legislation that is said to be needed to support TAV services in future state and local government elections.
The interim report sought legislative change only to prevent a TAV system failure from invalidating the results of an election. That recommendation carries through to the final report but is augmented by recommendations for three extra legislative changes.
The first of the extra changes sought would lengthen the time “between the close of candidate nominations and subsequent ballot paper draws” for upper and lower house elections.
The NSW Electoral Commission-led review is concerned the timeframe is tailored to a paper-based process and does not afford appropriate time to format a digital voting system.
“Lengthening this timeframe will provide more time to sufficiently prepare any future TAV systems with candidate information and to complete user testing,” the final report advises.
The second change sought is to how preferences are counted for NSW upper house elections.
The report notes that the current method does not “take full advantage of … digital scanning and counting technology” that is commonly used in other jurisdictions.
“Computational data entry and full electronic distribution of preferences is now considered best practice,” the report notes.
As to how this would aid the re-introduction of technology-assisted voting, the commission states that “a full distribution of preferences would assist in any future assessment of the materiality to the outcome of an election in the event of eligible electors not being to cast a vote due to technical performance of a TAV system or other disruption.”
The third legislative change would “rationalise the way parties, groups and candidates are displayed” on the upper house ballot paper.
The final report notes that lengthy paper-based ballots are hard to digitise; on internet-based channels, it can lead to endless scrolling and accusations of unfair preferential positioning for some candidates over others - although randomisation of the order could resolve the latter.
It’s worse for telephone-assisted voting: “It is very time consuming for telephone operators (and electors) to read out all candidates ‘below the line’ for telephone-assisted voters, adding to the logistics of that method of TAV.”
NSW had an online voting system called iVote but stopped using it after a technical glitch prevented some citizens from casting votes in a local government ballot, leading to court ordered re-votes in some areas.
Cyber security concerns elevated
The final report also elevates cyber threats as a reason to tread carefully on any reintroduction of TAV in NSW.
This is made clear upfront, with a new paragraph stating that “internet voting is a high-risk channel, facing a worsening cyber and misinformation threat environment involving state and criminal actors seeking to disrupt elections.”
“Moreover, the processes for verifying votes and other assurance steps are not generally understood by electors or political participants,” it states.’
The potential benefits of linking digital identity systems to TAV is far more qualified than in the interim report.
The interim report was more assured of the benefits: “Requiring electors to provide digital identity to use TAV platforms is a means of strengthening the integrity of the system. It provides greater assurance around the identity of an elector and limits the risk of system infiltration by bots or malicious third parties.”
The final report is more constrained in its language, however, adding the qualifiers ‘may’ and ‘could’ when touting the linkage benefits: “Requiring electors to provide digital identity to use TAV platforms could strengthen both the integrity of and trust in those platforms. It may provide greater assurance around the identity of an elector and limits the risk of system infiltration by bots or malicious third parties.”
It appears Vision Australia had some influence over the downplaying of digital ID’s potential benefits for TAV; the organisation was largely underwhelmed by MyGovID’s usability and take-up rates.
The final report largely carries through with the interim report’s recommendations that internet voting be scoped and costed with a view to a small-scale revival for the 2027 state election.
It also recommends some “limited-scale” digital voting kiosk trials for the 2028 local government elections.
However, the introduction of either would be dependent on finding an appropriate technology-based solution that is not too costly, can be configured in time, and is supported legislatively.
