Spotting Persuasion: A Low-cost Model for Persuasion Detection in Political Ads on Social Media

2025-03-18

AI Summary

This study addresses the automatic identification of persuasive text in social media political advertisements. We propose a lightweight BERT variant that integrates feature distillation with a task-adapted classification head, trained on a manually annotated corpus of Facebook political ads from the 2022 Australian federal election to ensure domain specificity. To our knowledge, this is the first work to jointly optimize model efficiency and real-world political ad analysis. Evaluated on SemEval 2023 Task 3 Subtask 3, our model achieves an F1-score of 78.4%, outperforming the baseline by 4.2 points, while maintaining a compact size of only 12 MB. Through interpretability analysis, we identify six prevalent persuasion strategies—including emotional arousal and moral labeling—enhancing transparency in political ad analysis. The approach provides an efficient, deployable solution for election interference monitoring in resource-constrained environments.

Technology Category

• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Machine Learning: Evaluation and Analysis
• Application Domains: Humanities & Computational Social Science

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

In the realm of political advertising, persuasion operates as a pivotal element within the broader framework of propaganda, exerting profound influences on public opinion and electoral outcomes. In this paper, we (1) introduce a lightweight model for persuasive text detection that achieves state-of-the-art performance in Subtask 3 of SemEval 2023 Task 3, while significantly reducing the computational resource requirements; and (2) leverage the proposed model to gain insights into political campaigning strategies on social media platforms by applying it to a real-world dataset we curated, consisting of Facebook political ads from the 2022 Australian Federal election campaign. Our study shows how subtleties can be found in persuasive political advertisements and presents a pragmatic approach to detect and analyze such strategies with limited resources, enhancing transparency in social media political campaigns.

Problem

• Detects persuasive text in political ads efficiently.
• Analyzes political campaigning strategies on social media.
• Enhances transparency in social media political campaigns.

Innovation

• Lightweight model for persuasive text detection
• Reduced computational resource requirements
• Applied to real-world Facebook political ads