Verifiability of electronic voting: between confidence and trust

When computing scientists speak about electronic voting, it is often in terms of trust. But there are two contradictory statements. First, they argue that it should not be necessary to trust e-voting systems, which would be the case if they are provably secure. Second, for an e-voting system to be successful, the public must trust it. When we unravel the confusing concept of trust, we find that there are two quite different meanings: relying on something that one does not understand and does not really choose (confidence), or relying on something that one does understand and has consciously chosen (trust). The distinction is due to the German sociologist Niklas Luhmann. In this contribution, we analyse how this distinction can help in analysing the controversies around electronic voting. It is argued that because of the controversy, paper voting and e-voting now tend to be seen as radically different alternatives, which require comparison and a conscious decision. Trustworthiness, as opposed to reliability only, has thereby become a major requirement of electronic voting systems, leading to the implementation of various verification options. This increasingly applies to other systems that handle sensitive data as well. We will discuss the various types of verifiability in electronic voting systems, and how these can contribute to trustworthiness of data processing in general.