By Dipesh Kaura, Country Director, India and SAARC, Securonix
During the recent United Nations Security Council Meeting at the UN General Assembly in New York, Anthropic and OpenAI CEOs warned that AI could threaten humanity without the application of necessary safeguards. If AI operates without human oversight, organizations risk undermining strategic priorities and long-term resilience. Human involvement in critical decision-making remains essential to keeping AI aligned with business objectives. Today, organizations across industries are increasingly adopting AI across business functions to enhance efficiency, strengthen decision-making, boost productivity, drive innovation, and deliver measurable business impact.
AI is also changing cybersecurity at a remarkable pace. This October, during Cybersecurity Awareness Month, we should note that AI has not only become a working member of the security team but also a tool in the attacker’s arsenal.
AI has entered the SOC
We are witnessing the growing demands of today’s cyber landscape, where the Security Operations Center (SOC) needs more than additional tools and requires a fundamental transformation in how security operations are managed. Security teams are overwhelmed by alert volumes, many of which are false, and are stretched thin across a continuously expanding attack surface. Costs are on the rise too, and the board is demanding proof, not just that threats are being stopped but whether the adopted security strategy is driving measurable value across the business.
CISOs face a dual mandate of defending against relentless, AI-powered threats while delivering outcomes the business and board can trust. AI-powered security can analyze large volumes of telemetry, freeing analysts to focus on higher-end tasks, complex investigations, threat hunting, incident strategy, risk decisions, and communicating with leadership. The goal of AI in security is not to remove people from the process, but to take the repetitive, high-volume work off their plates.
Human Oversight Should Be a Design Choice in AI
As organisations introduce Agentic AI into security operations, human oversight should not be treated as an afterthought. Furthermore, AI should empower human expertise and not replace it. Agentic AI has to be designed with analysts at the center, ensuring every action, insight, and automation enhances human judgment, providing intelligent support without compromising control. In this human-in-the-loop approach, the modular agents automate repetitive tasks, surface enriched insights, and accelerate decisions, but the analyst is always in control.
Be it tuning detection logic, approving response actions, or refining risk scoring, the analysts guide the loop, ensuring humans own the outcome with AI only supporting the work. Security teams should be able to understand what an AI system did, why it performed a particular action, what information influenced its decision, and how to intervene when necessary. Human oversight works best when it is built into how the technology is deployed, not added afterward.
AI Agents Require Transparency and Governance
The Human-in-the-loop approach ensures transparency, with analysts able to see what and why agents are doing. There are no black-box decisions, with actions being observable, auditable, and reversible.
These autonomous AI agents learn from analyst input while improving over time. Various Agents perform different tasks, such as frontline defense, enriching alerts with context, hunting threats, containing actions, detecting risky user behavior, and classifying and routing telemetry in real time. As
AI Agents interact with users, identities, applications, sensitive data, tools, and business processes, they need to be governed by security teams. These teams require visibility into the actions of these agents, including which resources they are accessing and whether changes in their behavior indicate misuse, policy violations, or compromise. Human-in-the-loop ensures flagging actions, providing context, and making it analysts to step in, make decisions, or override when required, while establishing trust, transparency, and control. The SOC should also be equipped with Governed AI Agent Detection and Response capabilities, establishing accountability and creating a foundation for responsible AI adoption.
The Future of Cybersecurity Will Be Human-Led and AI-Powered
Despite the increasing use of AI-powered security operations in SOCs, the fundamentals of cybersecurity remain critical. Strong identity security, effective access governance, employee awareness, and sound cyber hygiene continue to shape the first line of defense.
AI will certainly play a defining role in the future of cybersecurity. Agentic systems, intelligent assistants, and automated investigations are increasingly becoming part of the modern SOC. However, the objective should not simply be to build fully autonomous security operations. Instead, the future is likely to be shaped by a hybrid approach where AI brings speed, scale, and automation, while human expertise provides context, judgment, and accountability.
As cybersecurity becomes increasingly driven by machine intelligence, human oversight will remain essential to ensuring that AI is used responsibly, effectively, and in line with business priorities.