Detecting and Mitigating AI-Generated Cyberthreats

The rapid adoption of generative AI by adversaries is accelerating the creation of novel malware, automated scanners, and exploit campaigns, significantly reducing the time between vulnerability disclosure and active exploitation. Traditional signature-based detection and manually crafted honeypots struggle to keep pace with this evolving threat landscape.This presentation introduces a new, community-driven security tool designed to automatically detect, capture, and block emerging AI-generated cyberthreats and newly disclosed CVEs. Leveraging over a decade of real-world attack telemetry and threat data from the STINGAR platform, our AI model dynamically generates and deploys open-source honeypots that accurately emulate vulnerable services and configurations. These honeypots enable rapid identification of active exploitation attempts, provide high-fidelity attack intelligence, and support near-real-time defensive response across university and research networks. By automating honeypot creation and incorporating community feedback, the platform significantly reduces response time to zero-day and fast-moving threats while remaining transparent, extensible, and vendor-neutral. The session will cover system architecture, deployment models, and real-world use cases, and discuss how collaborative threat intelligence can strengthen defenses against AI-enabled adversaries.