Most security leaders report that current tools and skills are not fit to secure AI systems, according to Pentera’s AI and Adversarial Testing Benchmark Report 2026. The survey of 300 US CISOs found widespread limited visibility into AI usage, a primary skills gap, and heavy reliance on legacy controls rather than AI-specific tooling. #Pentera #AIandAdversarialTestingBenchmarkReport2026
Keypoints
- 67 percent of CISOs reported limited visibility into how AI is used across their organizations.
- Lack of internal AI security expertise was cited as the top obstacle by 50 percent of respondents.
- Organizations are extending legacy controls to protect AI, with 75 percent relying on traditional security tools.
- Only 11 percent reported having AI-specific security tools, while just 17 percent cited budget as a primary concern.
- The report advises building specialized skills and actively testing controls in environments where AI is operating.
Read More: https://thehackernews.com/2026/03/ai-is-everywhere-but-cisos-are-still.html