The Unpredictability Problem
The rapid adoption of artificial intelligence agents in enterprise environments has created a fundamental security challenge: AI applications are non-deterministic, meaning they do not behave according to fixed, predefined rules like traditional software. This unpredictability, combined with the immense potential blast radius of a compromised AI system and intense business pressure to deploy quickly, has left security teams struggling to keep pace. Niv Braun, co-founder and CEO of Noma Security, argues that legacy security models are inadequate for this new reality.
A Unified Approach to AI Protection
According to Braun, effective AI security must rest on two pillars: a flexible framework that can accommodate fast evolving technologies such as the Model Context Protocol (MCP), and deep contextualization that merges posture management, access controls, and runtime monitoring into a single, coherent signal. Without insight into what happens during runtime, it is impossible to provide accurate recommendations for configuration or access permissions. Braun advocates for a unified AI security platform over siloed point products, enabling teams to distinguish legitimate agent actions from genuine threats.
Source: Healthcareinfosecurity
