The AI Arms Race in Cyber: How Attackers Scale and How Defenders Are Adapting
Artificial intelligence is compressing the cybersecurity offense and defense cycle, changing how attackers scale and how defenders must adapt. Leo Taddeo, CEO and President of AppGate, presents a 1ArtificialIntelligence keynote on how AI enables attackers to generate phishing, spoofing, impersonation, and social engineering content at scale, accelerate vulnerability discovery at machine speed, and run agent-based workflows continuously. The session examines how trust manipulation becomes better and cheaper through personalization, multilingual lures, synthetic profiles, deepfakes, QR-based social engineering, and bot-driven trust building, while vulnerability discovery becomes faster, more automated, and less dependent on elite technical skill. Taddeo also outlines how defenders are adapting by disrupting impersonation earlier in the kill chain, reducing fraud operations overhead, using machine-learning decisioning and adaptive authentication, shrinking the attack surface, segmenting access, and placing crown-jewel systems behind explicit policy-based controls. A central leadership theme is the emergence of AI agents as a new risk surface, where autonomous, non-human actors with memory, tools, and system access require network-layer access control, microsegmentation, Zero Trust, and continuous verification to prevent over-privileged access, lateral movement, cascading failures, unintended actions, and data exposure.
