Anticipatory AI: Predicting Systemic Weaknesses in Democratic Disinformation Resilience

While Surjatmodjo et al. (2024) and Chan et al. (2023) document factors influencing resilience to disinformation (e.g., digital literacy, trust, platform design), most current work focuses on static or average-case analyses. Building on the "deviations from expectations" heuristic, this project proposes using AI to dynamically model and forecast resilience, using real-time data on disinformation exposure and public reactions. By training models on historical data—including instances where resilience was unexpectedly high or low—we can pinpoint systemic "blind spots" (e.g., demographic groups, platform types, or moments of heightened vulnerability). This approach differs from prior frameworks by not just mapping risk factors, but actively predicting where and when the system will fail or outperform. Such anticipatory modeling could inform rapid, targeted interventions, moving beyond broad, static policy recommendations toward agile, data-driven democratic defense.

References:

1. Information Pandemic: A Critical Review of Disinformation Spread on Social Media and Its Implications for State Resilience. Dwi Surjatmodjo, A. A. Unde, Hafied Cangara, Alem Febri Sonni (2024). The social science.
2. Distributed Discovery of News and Perceived Misinformation Exposure: A Cross-Continent Application of the Resilience to Online Disinformation Framework. Michael Chan, Dmitry Kuznetsov, Jingjing Yi, Francis L. F. Lee, Hsuan-Ting Chen (2023). The International Journal of Press/Politics.