Researchers have developed a computational method to analyze political delegitimization discourse (PDD) in Hebrew-language political speech. Their approach combines finetuned encoder models with decoder LLMs, achieving an F1 score of 0.74 for PDD detection. Analysis of Knesset speeches, Facebook posts, and news articles revealed a significant increase in PDD over three decades, with higher prevalence on social media and among right-leaning actors, particularly during election campaigns. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Demonstrates the potential for AI to analyze and understand complex political discourse, offering insights into democratic health.
RANK_REASON Academic paper presenting a novel computational method for analyzing political discourse. [lever_c_demoted from research: ic=1 ai=1.0]