Nahid Anwar – Cyber Security
Location: CCP Conf 368 or on Zoom
Title: Techniques for Detecting Election Fraud
National elections may be viewed as large-scale social experiments in which voters express their opinions through a ballot in a country that is divided into a disproportionately large number of voting districts or polling stations. The idea of free and fair elections is the foundation of a democratic society where an essential property is that every voter’s vote should be counted equally. Thus, elections are essential mechanisms for ensuring public accountability, transparency, and representation. Fraud and manipulation corrupt this electoral system and prevent voters’ voices from being heard. Fraud undermines the legitimacy of the democratic process when uncovered or even suspected and may lead to repression, violent unrest, and even civil war. Therefore, it is crucial to determine whether an electoral outcome reflects voter preferences or the result of manipulation. The credibility of an elected representative democracy depends on the integrity of the election system. There haven’t been many quantitative techniques for evaluating the integrity of a democratic election up until recently, other than in-person election observation. In this presentation, we discuss a few quantitative methodologies that have recently been developed in this area along with machine learning tools, which collectively form the nascent interdisciplinary field of election forensics for identifying polling locations at risk of election fraud and estimating the scope of potential electoral manipulation. In electoral forensics, anomalies in election results are looked for in order to better identify patterns indicative of electoral irregularities.