SandboxAQ released its inaugural AI Security Benchmark Report, revealing a significant disconnect between enterprise AI adoption and cybersecurity readiness. While 79% of organizations are already using AI in production environments, only 6% have implemented a comprehensive, AI-native security strategy, leaving the vast majority of enterprises vulnerable to threats they are not yet equipped to detect or mitigate. The report highlights widespread concern about the risks AI introduces, from model manipulation and data leakage to adversarial attacks and the misuse of non-human identities (NHIs). Yet despite growing anxiety among CISOs, only 28% of organizations have conducted a full AI-specific security assessment, and most are still relying on traditional, rule-based tools that were never designed to address dynamic, machine-speed systems. Only 6% of organizations have implemented AI-native security protections across both IT and AI systems. 74% of security leaders are highly concerned about AI-enhanced cyberattacks, and 69% are highly concerned about AI uncovering new vulnerabilities in their environments. Just 10% of companies have a dedicated AI security team; in most organizations, responsibility falls to traditional IT or security teams. Marc Manzano, General Manager of the Cybersecurity Group at SandboxAQ said, ” This report highlights a growing recognition among security leaders that defending against evolving threats requires new assumptions and approaches, not just new layers or patches to current tooling.”