ISCAP Proceedings: Abstract Presentation
How and When Does Generative AI Shift Cybersecurity Advantage?
David Yates
Bentley University
Abstract
Generative AI (GenAI) is transforming both cyber offense and defense, yet its net impact on relative advantage remains an open question. Rather than assuming GenAI inherently favors attackers or defenders, the present study investigates how GenAI shifts the underlying resource constraints through which cybersecurity outcomes are produced. We conduct a comparative analysis of fifteen documented attacker–defender episodes, structured across five attack families and three outcome states: attacker advantage, contested, and defender advantage. Each episode is coded for AI-driven resource shifts on both sides and evaluated based on remaining bottlenecks and their control. Our comparative analysis suggests a recurring mechanism: Generative AI most consistently alleviates constraints on expertise and time, while selectively relaxing barriers to scale, coordination, tactical judgment, and AI-accessible context. As these constraints diminish, the binding constraint migrates toward complementary resources required to convert AI-enabled capability into consequential effects — specifically access, infrastructure, privileged/local context, and operational control. Attacker advantage emerges when adversaries secure or bypass these remaining bottlenecks prior to defensive intervention. Defender advantage takes hold when defenders retain or reclaim them before consequential conversion. Contested outcomes arise when control is partial or shifts. The core contribution is a resource-bottleneck explanation of cybersecurity advantage: GenAI does not predetermine who wins, but redefines which resources are scarce and therefore which resources govern advantage.