Australian Electoral Commission Refuses To Allow Researchers To Check E-Voting Software
The fact that Techdirt has been writing about e-voting problems for sixteen years, and that the very first post on the topic had the headline “E-voting is Not Safe,” gives an indication of what a troubled area this is. Despite the evidence that stringent controls are still needed to avoid the risk of electoral fraud, some people seem naively to assume that e-voting is now a mature and safe technology that can be deployed without further thought.
In Australia, for example, e-voting is being used for the elections to the country’s Senate, but the Australian Electoral Commission (AEC) has refused to release the relevant software, despite a Senate motion and a freedom of information request. Being able to examine the code is a fundamental requirement, since there is no way of knowing what “black box” e-voting systems are doing with the votes that are entered. A story by the Australian Associated Press (AAP) explains why AEC is resisting:
The Australian Electoral Commission referred AAP to a decision by the Administrative Appeals Tribunal [AAT] in December 2015.
In that decision, relating to a freedom of information request, the tribunal found the release of the source code for the software known as Easycount would have the potential to diminish its commercial value.
“The tribunal is satisfied that the Easycount source code is a trade secret and is exempt from disclosure,” the AAT said.
Placing trade secrets above the public interest is a curious choice, to say the least. It seems particularly questionable given Australia’s recent experience with e-voting software problems:
When the ACT Electoral Commission released its counting code, researchers at Australian National University found three bugs which were subsequently fixed before an election.
When the Victorian Electoral Commission made its electronic voting protocol available to researchers in 2010, University of Melbourne researchers identified a security weakness which was then rectified before the state election.
As Techdirt readers well know, bugs are commonplace, and there’s no particular shame if some are found in a complex piece of software. But refusing to allow independent researchers to look for those bugs so that they can be fixed is inexcusable when the integrity of the democratic selection process is at stake.