Operating modelRelevance · medium2026-07-13
SANS Institute·5 min read
AI-driven security adoption is surging, but failures reveal a governance and skills gap
TechnologyFinancial ServicesEnergyPublic Sector
A July 2026 SANS Institute survey reports that AI use in cybersecurity jumped from 50% to 78% in a year, but AI-related failures and incidents also rose sharply, with most security teams holding governance responsibilities without formal audit frameworks or adequate workforce support.
Why it matters for leaders
As the technology AI/Data lead, you are likely being asked to embed AI into security operations while also owning parts of AI governance. This survey shows that without investing in evaluation, incident playbooks, and skills for both security and AI engineering teams, AI can increase rather than reduce risk.
Opportunity signal
Use the SANS findings to justify a joint AI–security program: shared evaluation pipelines for security-focused models, red-teaming of agentic workflows, and targeted upskilling for engineers and analysts on AI-specific failure modes.
mediumRisk signal
Deploying AI into security workflows without clear failure thresholds, human override processes, and monitoring can create blind spots that adversaries exploit and that regulators view as negligent.
high