70% of organizations give AI agents authority once requiring human approval

A survey of 1,100 IT and security decision-makers found 70% of organizations now let AI agents make decisions and take actions that required human sign-off a year ago. Three in four hit at least one AI-related security issue in the past year, and 62% deployed AI tools before setting up safeguards.

• 70% of respondents let AI agents take actions that once needed human approval
• 75% of organizations saw at least one AI-related security issue or warning sign in a year
• 84% claim visibility into AI data access, but only 49% can audit every agent action
• 62% deployed AI tools before putting proper safeguards in place