The First Mass Protest on Threads: Multimodal Mobilization and AI-Generated Visuals in Taiwan's Bluebird Movement

TL;DR

The paper investigates how Taiwan's Bluebird Movement mobilized on Threads, a platform with a unique national audience and rapid growth. It combines a large multi-modal dataset (62,321 posts and 21,572 images) with zero-shot LLM labeling and CatBoost/SHAP analyses to examine how algorithmic exposure and user engagement diverge across partisan lines, and how textual and AI-generated visual strategies shape virality. Key findings include a pronounced exposure-engagement gap favoring anti-DPP content in views but anti-KMT and pro-DPP content in engagement, text that emphasizes commemorations and personal testimony as virality drivers, and AI-generated imagery that both mobilizes supporters and enables partisan attacks within a kawaii aesthetic. The study demonstrates how Threads functions as both an amplifier and a selective moderator of democratic contention, and introduces the concept of kawaii toxicity to describe how cute AI visuals cloak toxicity in political discourse, with implications for digital mobilization and platform governance.

Abstract

The 2024 Bluebird Movement in Taiwan marked one of the largest youth-led protests in the country's democratic history, mobilizing over 100,000 demonstrators in response to parliamentary reforms. Unlike the 2014 Sunflower Movement, Bluebird unfolded within a transformed digital environment dominated by Threads, Meta's new microblogging platform that$\unicode{x2013}$uniquely$\unicode{x2013}$draws 24% of its global traffic from Taiwan. Leveraging a dataset of 62,321 posts and 21,572 images, this study analyzes how protest communication developed across textual and visual modalities. We combine LLM zero-shot annotation, gradient-boosting trees, and SHAP explainers to disambiguate the supply and demand of attention. Results reveal three dynamics: (1) partisan asymmetries between algorithmic exposure and user endorsement, with anti-DPP content surfaced more widely but anti-KMT and pro-DPP content more actively recirculated; (2) textual repertoires centered on commemorations, personal testimonies, and calls to action as key drivers of virality; and (3) a bifurcation in visual strategies, where human photographs concentrated exposure and discussion, while AI-generated animal and plant symbols circulated as mobilization tools and partisan attacks. These findings demonstrate how Threads functioned as both an amplifier and filter of democratic contention, extending theories of emotional and visual contagion by showing how generative AI reshapes symbolic repertoires in contemporary protest through what we term kawaii toxicity$\unicode{x2013}$political attacks cloaked in aesthetics of cuteness.

Figures (6)

• Figure 1: Annotated timeline of the Bluebird Movement between 05/01/2024 to 06/01/2025.
• Figure 2: Textual proportion of Bluebird Threads content, with the (a) partisan framing and (b) topical mentions within all posts.
• Figure 3: Top four partisan stances weighted by exposure (views) and engagement (reposts, likes, and replies).
• Figure 4: SHAP explainers of the CatBoost Regressor that predicts engagement (likes) based on a) textual and b) visual features.
• Figure 5: Attention funnel for visual framing categories across exposure and engagement metrics (views, reposts, likes, and replies).