Top cybersecurity executives at RSAC 2025 emphasized that artificial intelligence's role in security has evolved rapidly, shifting from experimental use to core operational functions, CRN reports.
Leaders from CrowdStrike, SentinelOne, Palo Alto Networks, Wiz, and others agreed that the next frontier lies in agentic AI, or intelligent systems that act autonomously within security programs. CrowdStrike's Daniel Bernard framed the pivotal question: "Can I trust an AI agent to operate a security program or part of a security program for me?" As generative AI helps both defenders and attackers, industry figures warned of growing threats from AI-powered attacks and stressed the urgency of adapting defenses. Palo Alto's Lee Klarich pointed to a 300% rise in daily cyberattacks, while others, like Wiz's Ami Luttwak and SentinelOne's Tomer Weingarten, cautioned that AI security must stay agile as capabilities evolve. Executives underscored the need to move from blocking AI tools to enabling secure deployment, with Proofpoint's Sumit Dhawan predicting AI will act like "virtual humans," which will require their own security frameworks.
Leaders from CrowdStrike, SentinelOne, Palo Alto Networks, Wiz, and others agreed that the next frontier lies in agentic AI, or intelligent systems that act autonomously within security programs. CrowdStrike's Daniel Bernard framed the pivotal question: "Can I trust an AI agent to operate a security program or part of a security program for me?" As generative AI helps both defenders and attackers, industry figures warned of growing threats from AI-powered attacks and stressed the urgency of adapting defenses. Palo Alto's Lee Klarich pointed to a 300% rise in daily cyberattacks, while others, like Wiz's Ami Luttwak and SentinelOne's Tomer Weingarten, cautioned that AI security must stay agile as capabilities evolve. Executives underscored the need to move from blocking AI tools to enabling secure deployment, with Proofpoint's Sumit Dhawan predicting AI will act like "virtual humans," which will require their own security frameworks.
