Security Report · raw.githubusercontent.com

Pentera Survey Finds Widespread AI Use and Limited Security Visibility at U.S. Enterprises

Oct 21, 2026 · Original Report (PDF) ↗

Report discovered via Awesome Annual Security Reports.

Summary

Pentera’s survey of 300 U.S. security leaders found widespread enterprise AI use but limited visibility and few dedicated AI security budgets. Respondents also reported frequent unauthorized access and plans to increase security and pentesting spending.

Key points

The survey was commissioned by Pentera and conducted by Global Surveyz in December 2025; it covers self-reported responses from 300 U.S. CISOs and senior security executives at organizations with at least 3,000 employees.
All respondents reported some enterprise AI use, but 66% said visibility into AI deployments was limited and 33% expected unmanaged AI use despite reporting good visibility.
Only 11% reported using tools specifically designed to protect AI systems; 75% said they rely on traditional security tools, and 78% fund AI security from existing budgets.
Respondents cited lack of internal expertise (50%) and limited visibility (48%) as leading AI security challenges; 44% said AI security lags behind their existing IT security controls.
At least 75% reported that attackers had gained unauthorized access to their environment in the preceding 24 months. The report does not provide independent incident verification.
For 2026, 66% planned to increase overall security spending and nearly 70% planned to increase pentesting budgets; 21% expected to create a dedicated AI security budget within 12 months.

Article Details

Publisher: Pentera
Report Period: Survey conducted in December 2025; report covers 2026
Scope: Self-reported survey of U.S. enterprises on AI security, cybersecurity posture, and security spending.
Sample Size: 300 U.S. enterprises; respondents were CISOs and senior security executives.
Key Statistics: At least 75% of surveyed enterprises reported that attackers gained unauthorized access to their environment in the past 24 months.
66% of CISOs reported limited visibility into AI usage; 1% reported no visibility.
99.7% of enterprise CISOs reported that securing their AI ecosystem was driving some level of increased cybersecurity spending in 2026.
CISOs managed an average of 47 security solutions; 40% of organizations operated 51 or more.
The weighted average cybersecurity budget was approximately $2.48 million, with approximately $300,000 allocated to pentesting.
Recommendations: Treat AI as a governed enterprise asset rather than allowing ad hoc adoption.
Increase the cadence and scope of adversarial testing to validate real-world exposure.
Evaluate how AI expands or accelerates existing attack paths rather than treating it as a standalone risk domain.
Prioritize team upskilling and stronger validation practices before adding more point solutions.

Vendors

Pentera- [About Pentera](#about-pentera)

Industries