The Unpredictability Challenge
The rapid adoption of artificial intelligence agents and applications is creating a significant security challenge for enterprises. Unlike traditional deterministic software that follows predefined rules, AI systems behave unpredictably, making it difficult for security teams to keep pace. Niv Braun, CEO of Noma Security, explains that this non deterministic nature vastly expands the potential blast radius of any security incident. The pressure to deploy AI quickly often leaves security teams scrambling to catch up, creating a perfect storm for vulnerabilities.
A Framework for AI Security
To address these challenges, Braun advocates for a security approach built on two core principles: a holistic framework flexible enough to incorporate rapidly evolving technologies like the Model Context Protocol (MCP), and deep contextualization that unifies posture management, access controls, and runtime monitoring into a single, cohesive signal. Without visibility into runtime behavior, security teams cannot provide accurate recommendations for configuring AI systems or determining appropriate access levels for agents. This comprehensive view is essential for distinguishing legitimate agent actions from those that pose genuine risk.
Impact and Scope
The interview, conducted at RSAC Conference 2026, highlights the growing need for early partnerships between AI providers and security vendors to enable secure by design capabilities. Braun emphasizes that a unified AI security platform outperforms siloed point products by connecting security insights across the entire AI lifecycle. As organizations race to integrate AI into their operations, the ability to maintain context and control over these systems will be critical to preventing widespread security failures.
Source: Healthcareinfosecurity
