Securing Autonomous AI: Governance, MCP & Agentic AI Risks | CISO Roundtable March 2026

AI agents are no longer just assisting they are beginning to act independently, call tools, access data, and collaborate across systems. In this March 2026 Interactive CISO Roundtable, security leaders explore one of the most important emerging challenges: 👉 How do you secure autonomous AI systems? With the rise of Agentic AI and frameworks like Model Context Protocol (MCP), organizations must rethink governance, security controls, and observability for systems that can operate with increasing independence. This session brings together experienced CISOs and AI leaders to share practical insights on securing agentic systems in real-world environments. 💡 Topics Covered • Evolving threat models for agentic systems (prompt injection, tool abuse, data exfiltration) • Security advantages and risks of MCP as a standardized interface • Enforcing policies across agent-to-tool communication • Observability, logging, and AI audit trails for monitoring AI agents • Applying zero-trust principles to autonomous AI systems 🎙 Speakers Arsalan Khan — CISO, Graphiant Randolph Barr — CISO, Cequence Security Maneesh Sahu — Founder, Ferentin 🎤 Host: Naveen S. Bisht — CEO, Akitra If you’re a CISO, GRC leader, AI architect, or security engineer, this discussion will help you understand how to design secure, governed, and observable AI-driven systems. 👉 Subscribe to Akitra for more expert discussions on Agentic AI, AI security, and modern cybersecurity leadership. #CISORoundtable #AISecurity #AgenticAI #CyberSecurity #AIinSecurity #ZeroTrust #SecurityLeadership #Akitra