Internet voting has the potential to ease voter participation and provide a high-tech upgrade to traditional polling methods. Unfortunately, it also raises some of the most difficult challenges in computer security, due to the need to safeguard election servers and voters' computers against powerful attackers, while simultaneously protecting the secret ballot. How well can election technology defend against modern security threats? To find out, colleagues and I performed in-depth security evaluations of Internet voting systems used in the U.S. and around the world. We found staggering gaps in system designs and operational procedures—problems that would allow attackers to change votes, compromise privacy, disrupt returns, or cast doubt on election results. These case studies illustrate the practical obstacles to securing Internet voting and carry lessons for any locality considering adopting such systems.