Survey Surfaces Greater CISO Appreciation for Scope of AI Threat

**Executive summary:** A Censuswide survey for Vorlon of 500 US CISOs reports widespread concern and observed incidents involving SaaS and AI tools—31% reported unauthorized data exfiltration between SaaS and AI, 87% lack visibility into data exchanged with AI, and most expect increased SaaS/AI security budgets—while noting a gap between perceived controls and actual exposures, especially around OAuth/API token compromise and emerging AI agent risks.