Disinformation in Democratic Deliberation: Assessing Patterns of Disinformation in European Parliament Debates Using Large Language Models

The rise of disinformation in politics in recent years, driven by the electoral success of populist and radical political parties, is often considered a threat to liberal democracy. Research so far has primarily focused on false or misleading claims circulating on social media and during election campaigns. We therefore know little about how disinformation manifests in everyday parliamentary debate - a core arena of democratic decision-making. This paper addresses this gap by analysing patterns of disinformation in the European Parliament, using an original corpus of debate records from 2009 to the present and a multi-step large language model pipeline to identify and classify factual accuracy and narrative content in legislative speechmaking. Preliminary findings reveal variation in disinformation across topics, time, and party ideology. The study advances our understanding of disinformation within parliamentary arenas as well as the role of political elites in its dissemination.