Designing for the Unlikely: Mitigation Against Rare, High-Impact Threats

Cybersecurity strategies are typically built on risk-based frameworks that prioritize threats according to their perceived likelihood, impact, and the cost of mitigation. Although effective against frequent and high probability risks, this approach systematically underestimates rare but high-impact events. Such events, though statistically unlikely, can produce catastrophic disruption, as illustrated by the 9/11 attacks or the COVID-19 pandemic. Preventing these events would have required sustained investments that seemed economically or socially unjustifiable until after the disaster struck. This tension between efficiency and resilience exposes a systemic weakness: defenders often prioritize short-term risks over resilience for rare adversarial scenarios. This paper argues that it is feasible and necessary to develop novel countermeasures that address rare and high-impact risks without imposing prohibitive costs. Using autonomous drone-based border monitoring as a case study, we examine how attackers can exploit the system by injecting fake events. We propose mitigation strategies that introduce modest inefficiencies, such as random perturbations or suboptimal path choices, which substantially reduce susceptibility to manipulation while adding minimal overhead. We argue for a paradigm shift: from optimizing information systems exclusively for efficiency to designing resilient systems that can withstand rare but high-impact events.