Cybersecurity and Elections

Claims of election fraud and threats to election security are frequent in contemporary U.S. political discourse, but to what extent are elections actually at risk? J. Alex Halderman, Professor of Computer Science and Engineering, is a cybersecurity expert whose research investigates threats to U.S. elections in an effort to answer precisely that question. Halderman engages in security testing, in which he evaluates the security of election systems through probing for vulnerabilities by simulating an attacker’s approach. He has recently submitted expert reports on critical insecurities in the state of Georgia’s polling equipment and acted as an expert witness in a hearing investigating whether Georgia’s elections are susceptible to manipulation. His testimony received national coverage: “Inside Georgia’s effort to secure voting machines as experts raise concern,” PBS News Hour (9/27/2024)

In addition to identifying risks, Halderman’s research also proposes cybersecurity solutions to help improve election security. His recent work has developed an algorithm for designing “test decks” for voting machines. These “test decks” detect any misconfigurations that could cause votes to be counted incorrectly. The system has already been piloted in Michigan elections as a low-cost way to safeguard election security. In other work aimed at improving elections, Halderman has helped develop an approach for using machine learning to improve ballot scanning accuracy. His computer vision techniques have allowed him to develop an algorithm for manipulating images of ballots to show that secure elections require counting from physical ballots rather than ballot images. You can find more of his work on election security (and other topics) at his website.