Skip to content
InfoNews

The Path to AI Autonomy in the SOC

Published
Record updated
View JSON

Summary

An a16z survey of CISOs at organizations with more than 1,000 employees found that 36% are comfortable letting AI agents act autonomously on low-risk security actions today, and 60% expect that level in 12 months. The article argues that AI autonomy in the SOC should be set per workflow stage and per action rather than as a single slider. It maps four stages (evidence collection, analysis and verdict, response recommendation, response execution) and draws on the Parasuraman, Sheridan and Wickens framework and a five-level agent autonomy scale from University of Washington researchers.