Large language models help represent voter views and test party positions
Large Language Models and Augmented Democracy
Computers and SocietyComputation and LanguageCryptography and Security
Summary
Deciding what people and political parties really want is tricky, especially when lots of options exist. This research explores using AI models called digital twins, which act like virtual versions of people or parties to predict their policy preferences. The authors tested these AI twins with data from Brazil and Switzerland to see if they capture opinions well and reflect party differences better than official statements. They also looked at how these AI systems can be tricked or influenced and developed ways to detect and block such attacks. Their work helps show how AI can assist democratic discussions while being careful about accuracy and manipulation.
What this means in practice
- •For political campaign teams: Use AI digital twins to predict voter policy preferences and tailor campaign messaging more accurately.
- •For government policy analysts: Model political party positions with AI to understand the range of intra-party views beyond official statements for better policy planning.
Authors
Jairo Gudiño-Rosero
Abstract
Artificial intelligence enables computational agents to represent political preferences and take part in collective decision-making. In this thesis, I investigate the opportunities and challenges of digital twins (DTs) based on Large Language Models (LLMs) as intermediaries in augmented democracy, focusing on individual preference representation, collective representation of political organizations, and the vulnerability of those representations to attackers. First, using data from an online experiment in Brazil, I examine whether personalized DTs can predict citizens' preferences for unseen policy proposals. Second, I extend the DT framework from individuals to political organizations. Using Swiss parliamentary data, I build topic-specific knowledge graphs from lawmakers' legislative records and connect them to LLM-based lawmaker agents, which are organized into party-level DTs representing collective positions. Agentic deliberation among these agents tests whether aggregated party representations capture a broader range of intra-party perspectives than official party communications. Finally, I study the vulnerability and robustness of LLM-mediated deliberation against prompt-injection attacks that amplify viewpoints, suppress opinions, or redirect consensus. Using data from a 2023 deliberative experiment in the United Kingdom, I analyze how attack effectiveness varies with the distribution of opinions and rhetorical strategies, and evaluate a pipeline combining injection detection, structured opinion representations, and reinforcement learning to improve resistance. These findings characterize the opportunities and challenges of LLM-based digital twins in augmented democracy, stressing accurate preference representation, faithful aggregation, and robustness to strategic interaction.