How Much Control Should AI Get? Inside the SOC Autonomy Question

The SOC Autonomy Debate

Security operations centers are increasingly turning to AI to handle the growing volume of alerts and threats. The question of how much control these systems should have is becoming a central concern for security teams.

The New Stack's Carly Page explores this topic in an article published on October 6, 2026, examining the tension between AI automation and human oversight in SOC environments.

Balancing Automation and Oversight

As AI systems become more capable of detecting and responding to security incidents, organizations face difficult decisions about where to draw the line between autonomous action and human-controlled processes.

The article appears in The New Stack's AI Operations and Security sections, reflecting its relevance to both communities. The publication covers topics including AI engineering, security, and operations.

Industry Context

The article arrives amid broader industry conversations about AI agent autonomy. Other recent coverage on The New Stack has addressed AI agents in enterprise settings, including pieces on OpenClaw landing in the enterprise with OpenAI, Nvidia and Red Hat on board, and Dynatrace wanting its AI agents to fix problems instead of adding alerts.

Security-focused coverage has also examined how AI is speeding up exploits and how vulnerability spreadsheets cannot keep up, highlighting the pressure on SOC teams to adopt more automated approaches.

The Human Factor

The core question raised by the article is not whether AI can perform security tasks, but how much decision-making authority it should hold. This touches on fundamental issues of accountability, trust, and the role of human judgment in security operations.

As organizations deploy more AI-driven security tools, the SOC autonomy question will likely remain a key topic for security professionals and the vendors building these systems.

Sources: Thenewstack