AI-Powered SOCs: Automate Smarter, Defend Better

AI-driven SOCs promise efficiency gains, but how much automation is too much? While AI-powered security tools enhance SIEM, SOAR, and threat detection, adversaries are exploiting automation blind spots, manipulating AI models, and evading AI-driven defenses. This session dissects real-world adversarial AI attacks, including a case study on Microsoft Copilot’s AI vulnerabilities—where attackers exploited indirect prompt injection and adversarial inputs to manipulate security workflows. Attendees will learn how AI can be both a force multiplier and a security risk, gaining: - An understanding of how attackers bypass AI-powered security automation - A breakdown of adversarial AI tactics and AI-specific SOC vulnerabilities - A hybrid AI-human SOC model that reduces false positives while maintaining resilience This session offers a practical roadmap to securely integrate AI in SOC operations without increasing attack surfaces.