Cyberattacks Increasingly Caused By Unchecked AI Agents
SC Media, Tuesday, April 21st, 2026
65% of enterprises experienced AI agent-related cybersecurity incidents due to poor governance and visibility.
According to research by Cloud Security Alliance and Token Security, the lack of strategy around AI agent governance is significantly increasing cybersecurity risks for organizations.
The study found that 65% of enterprises experienced at least one AI agent-related cybersecurity incident in the past year, with 82% discovering undiscovered agents they weren't aware of. While over half of respondents reported confidence in their AI agent visibility, forgotten and improperly decommissioned agents pose substantial risks that can affect core business operations.
The CSA recommends organizations implement strategies for maintaining agent visibility, applying consistent lifecycle governance, setting operational boundaries, and requiring human approval for critical decisions as AI agents gain greater autonomy.