AI is Everywhere, But CISOs are Still Securing It with Yesterday's Skills and Tools, Study Finds

Executive Summary: Pentera's AI and Adversarial Testing Benchmark Report 2026, based on a survey of 300 US CISOs and senior security leaders, finds that organizations struggle to secure AI systems due to limited visibility into AI usage (67%), insufficient internal AI security expertise (50%), and heavy reliance on legacy security controls while only 11% use AI-specific tools; the report concludes that skills and validation practices, not budgets, are the primary gaps to address as AI is integrated into enterprise infrastructure.