Signal New taxonomy documents non-adversarial failure modes in optical-scan voting systems

Summary

Researchers introduced a taxonomy of non-adversarial failure modes in optical-scan voting systems and their supporting ecosystem of people, processes and technology, illustrated with documented real-world incidents. While a substantial body of prior work has examined adversarial threats to voting systems, the paper notes many jurisdictions have not fully internalized the possibility of ordinary tabulator error, and that stakeholders often find hypothetical attack scenarios less convincing than real-world accounts of equipment and procedural failures. The taxonomy organizes failures into four categories: recording votes on paper, reading votes from paper, combining votes into a reported outcome, and testing/verification. The researchers map common verification mechanisms against this taxonomy and identify gaps that no paper-based audit can detect or correct, most notably failures compromising the trustworthiness of the paper trail itself — such as giving voters the wrong ballot style, using ballot-marking devices improperly, or failing to keep voted ballots secure and organized.