Quality Standards of Voting Advice Applications

In May 2013, some 30 political scientists from across Europe met in Lausanne, Switzerland, for a workshop to discuss a wide range of aspects of Voting Advice Applications (VAAs). The workshop focused mainly on VAAs’ impact on parties and voters, but one of the outcomes was also the so-called Lausanne Declaration (Garzia & Marschall, 2014). This document represents the first and so far only attempt to establish a framework for quality standards for VAAs. In addition to actual design aspects (questionnaire design, matching algorithm, usability, etc.), the declaration also covers ethical aspects relating to organizational and operational areas of VAAs. Twelve years after its publication, the Lausanne Declaration appears largely outdated. The spread and use of VAAs have increased significantly since 2014, making a strong focus on quality and ethical aspects of these tools more important than ever. At the same time, both the political environment (including democratic backsliding and political polarization) as well as the technical environment (such as AI, LLMs, and chatbots) have changed and evolved significantly. As a result, the Lausanne Declaration no longer meets today's requirements and needs to be modernized. In this paper, we present, based on a broad literature review, an initial assessment of which aspects should be considered in a contemporary framework on the quality of VAAs, which problems and challenges VAAs face if they want to meet current democratic requirements, and which other ethical questions arise in connection with their use. The paper will not provide definite answers as to what requirements VAAs must meet in order to be considered as high quality. Rather, it is intended to provide a starting point for a broad discussion on this topic within both the VAA research community and among VAA providers.