ChatGPT as a voting application in direct democracy

Abstract

The study examines the potential role of ChatGPT as a tool for popular voting. It assesses ChatGPT's positions on four voting objects (three initiatives and one referendum) by simulating various Swiss voter profiles (neutral, centrist, left, right, progressist, and conservative) and comparing these to its default stance. Additionally, ChatGPT's responses to arguments for and against each object were analyzed for consistency with its voting patterns. The study compared ChatGPT's decisions to real voting outcomes, fixing the temperature and collecting 100 answers per question to ensure reliability. Results showed that voting decisions varied by model version (GPT-3.5 or GPT-4) and language (French or German), indicating cultural and model iteration influences. ChatGPT's decisions did not consistently align with a left-progressist stance, highlighting the impact of language, model versions, and contextual factors on political interpretations.